You are working directly inside the existing production repository for:

Your task is to audit, research, architect, implement, connect, validate, and report a complete AWS AI/ML + Generative AI + Agentic AI topical cluster end-to-end.

This is not just an Amazon Bedrock cluster.

The final implementation must comprehensively cover the AWS AI/ML ecosystem, including:

Amazon Bedrock
Amazon Bedrock AgentCore
Amazon Nova
Amazon SageMaker
Amazon SageMaker AI
SageMaker Unified Studio
SageMaker Lakehouse
SageMaker Catalog / Data and AI Governance
AWS MLOps
AWS RAG / vector search
AWS AI data services
AWS purpose-built AI services
AWS AI infrastructure
AWS AI security/governance
AWS AI observability
AWS AI DevOps / IaC
AWS agent frameworks
AWS production troubleshooting
AWS AI/ML job support
AWS AI/ML interview support
AWS AI/ML profile positioning
AWS AI/ML candidate marketing
AWS AI/ML interview scheduling
AWS AI/ML country pages
AWS AI/ML city pages
AWS AI/ML industry pages
AWS AI/ML guides
AWS AI/ML comparison pages

The final result must be an interconnected SEO + user-navigation + commercial funnel, not a collection of disconnected pages.

1. ABSOLUTE SAFETY RULE — DO NOT BREAK EXISTING SITE

This implementation is ADDITIVE-FIRST.

Do NOT:

delete an existing page
remove an existing route
rename an existing URL
create a duplicate URL
overwrite an existing working cluster
remove existing keywords
remove existing internal links
remove schema
remove existing content merely to simplify implementation
change branding
change company name
change phone number
change WhatsApp number
change contact information
change pricing
alter unrelated services
redesign the website
change global typography
change global colors
change header/footer unnecessarily
change homepage primary intent
break responsive behavior
break sitemap generation
break robots.txt
break llms.txt
break llms-full.txt
break structured data
break build/tests
create orphan pages
create duplicate pages targeting identical intent

If a proposed page already exists:

KEEP THE EXISTING URL.

Enhance it only where needed and connect it into the new cluster.

The repository is the source of truth.

2. PHASE 1 — FULL REPOSITORY AUDIT BEFORE WRITING ANY CONTENT

Do NOT start generating pages yet.

First inspect the entire repository and understand how this site is built.

Audit:

framework
routing architecture
content architecture
page templates
dynamic routes
static routes
reusable components
page generation utilities
SEO metadata system
canonical URL handling
JSON-LD
breadcrumbs
navigation
footer links
technologies page
services page
locations pages
knowledge-base pages
blog
interview pages
candidate marketing pages
job support pages
proxy interview pages
get-interview-scheduled pages
profile positioning pages
country clusters
city clusters
industry clusters
comparison pages
sitemap generation
robots.txt
llms.txt
llms-full.txt
RSS/feed if present
internal-link components
related-page components
CTA components
phone/WhatsApp components
tests
build scripts
linting
route validation

Study successful existing large clusters, especially:

AI/ML
Angular
.NET
Workday
UiPath
country/location clusters
candidate marketing cluster
interview support cluster

Determine which patterns are reusable.

Create an internal inventory of all relevant URLs classified as:

EXISTING
PARTIALLY COVERED
MISSING
OVERLAPPING
DUPLICATE RISK
STALE
LEGACY
NEEDS INTERNAL LINKING

Find every existing URL containing or strongly related to:

aws
amazon
bedrock
sagemaker
genai
generative-ai
ai-ml
mlops
rag
agentic-ai
agentcore
nova
opensearch
textract
rekognition
transcribe
polly
lex
comprehend
personalize
trainium
inferentia
cloudwatch
terraform
kubernetes
python
langchain
langgraph
vector
interview
candidate-marketing
profile-positioning

Do not create anything until this audit is complete.

3. PHASE 2 — AUGUST 2026 AWS FRESHNESS AUDIT

Before writing any AWS AI/ML content, research the latest available AWS product state through August 2026.

This is mandatory.

Primary sources should be official AWS sources:

AWS product pages
AWS Documentation
AWS What's New
AWS service release notes
AWS Architecture Center
AWS Prescriptive Guidance
AWS Decision Guides
AWS FAQs
official AWS model documentation
official AWS regional/service availability documentation where needed

Do not rely on stale 2024/2025 assumptions.

For every service or feature being indexed, verify:

current product name
current branding
current service status
current APIs
current SDK terminology
current capabilities
current supported models
current supported integrations
current deployment options
current inference options
current supported vector stores
current agent architecture
current security capabilities
current observability capabilities
current regions where relevant
new August 2026 capabilities
recently changed features
recently renamed services
deprecated capabilities
legacy capabilities
features unavailable to new customers
current recommended architecture

Every page created or materially updated during this project must reflect the latest verified state available through August 2026.

If content is version-sensitive, verify it dynamically before publishing.

Do not fake freshness by simply adding "2026" everywhere.

Freshness must come from accurate technical content.

Where the site supports update dates, use the real modification date and appropriate dateModified.

4. MASTER SITE ARCHITECTURE

Build a major parent hub around:

AWS AI/ML & Generative AI

The hierarchy should conceptually be:

AWS AI/ML ├── Amazon Bedrock ├── Amazon Bedrock AgentCore ├── Amazon Nova ├── Amazon SageMaker ├── Amazon SageMaker AI ├── AWS MLOps ├── AWS RAG / Vector Search ├── AWS AI Data ├── AWS Agentic AI ├── AWS Purpose-Built AI Services ├── AWS AI Infrastructure ├── AWS AI Application Architecture ├── AWS AI Security & Governance ├── AWS AI Observability ├── AWS AI DevOps / IaC ├── AWS AI/ML Interview Support ├── AWS AI/ML Career Services ├── AWS AI/ML Locations └── AWS AI/ML Knowledge Base

Amazon Bedrock should be the strongest Generative AI commercial hub.

SageMaker / SageMaker AI should be the strongest ML/MLOps hub.

AgentCore should be the strongest agent-production hub.

Do not create a technically incorrect taxonomy simply for SEO.

5. AMAZON BEDROCK CLUSTER

Audit and build complete coverage around current Amazon Bedrock capabilities.

Core Bedrock

Cover relevant distinct intents around:

Amazon Bedrock
AWS Bedrock
Bedrock job support
Bedrock production support
Bedrock live project support
Bedrock project onboarding
Bedrock implementation support
Bedrock architecture support
enterprise Bedrock
Bedrock application development
Bedrock troubleshooting

Reuse existing pages such as /aws-bedrock-job-support/ if already present.

Do not duplicate them.

Bedrock Runtime / APIs

Research and cover:

Bedrock Runtime
Converse API
ConverseStream
InvokeModel
InvokeModelWithResponseStream
synchronous inference
streaming inference
batch inference where current
inference request/response design
token handling
retry/backoff
quotas
SDK integration

Bedrock Inference

Cover:

on-demand inference
provisioned throughput
inference profiles
application inference profiles
cross-region inference
intelligent prompt routing
latency optimization
throughput optimization
cost optimization
throttling
quotas
concurrency

Foundation Models

Use current AWS model catalog.

Potential durable families:

Amazon Nova
Anthropic Claude
Meta Llama
Mistral
Cohere
embedding models
multimodal models
image/video models where current

Avoid creating fragile pages for every temporary minor model version unless search intent warrants it.

Prefer durable family/provider pages.

Bedrock Customization

Cover current supported capabilities:

fine-tuning
custom models
continued pre-training where current
reinforcement fine-tuning where current
distillation
model import
model customization
model evaluation
model selection
evaluation datasets
LLM-as-a-Judge where current

Prompt Layer

Cover:

Bedrock Prompt Management
prompt templates
prompt versioning
prompt optimization
system prompts
structured output patterns
prompt evaluation

Bedrock Flows

Cover:

Bedrock Flows
workflow orchestration
conditional flow logic
prompt/model nodes
Lambda integration
Knowledge Base integration
debugging
production workflows

6. BEDROCK KNOWLEDGE BASES / RAG CLUSTER

Build a deep dedicated cluster around:

Bedrock Knowledge Bases
Bedrock RAG
enterprise RAG
Knowledge Base architecture
ingestion
synchronization
data sources
Retrieve
RetrieveAndGenerate
chunking
semantic chunking
fixed-size chunking
hierarchical chunking where current
metadata filtering
reranking
query rewriting
query decomposition where current
structured retrieval
multimodal retrieval
embeddings
vector search
citations
evaluation
RAG quality
hallucination reduction
retrieval quality
latency
security
production debugging

Connect with current supported vector/search/data systems such as appropriate:

OpenSearch Service
OpenSearch Serverless
Aurora PostgreSQL
pgvector
S3
Neptune / graph systems
other currently supported vector stores

Research official support before publishing.

Cross-link to existing generic:

RAG
vector database
LangChain
LangGraph
AI/ML
GenAI
embeddings

pages rather than duplicating their intent.

7. BEDROCK GUARDRAILS / RESPONSIBLE AI

Build complete coverage around current Bedrock Guardrails capabilities.

Research:

content filters
denied topics
sensitive information
PII protection
word filters
contextual grounding
hallucination mitigation
automated reasoning where current
prompt-attack protection where current
guardrail evaluation
Guardrails API
applying Guardrails independently where current
responsible AI
enterprise governance
security architecture

Create standalone pages only when distinct search intent exists.

Otherwise group subfeatures into a strong Guardrails hub.

8. BEDROCK DATA AUTOMATION

Research current August 2026 Bedrock Data Automation capabilities and cover:

Bedrock Data Automation
intelligent document processing
document extraction
image understanding
audio understanding
video understanding
multimodal extraction
standard output
custom output
blueprints
visual grounding
confidence scores
Data Automation Library
custom vocabulary
Knowledge Bases integration
production troubleshooting

Evaluate industry-specific intent such as:

healthcare document processing
insurance documents
financial documents
claims
forms

Only create vertical pages where technically differentiated.

9. BEDROCK AGENTCORE — FULL MAJOR CLUSTER

Research current AgentCore architecture through August 2026.

Create a major AgentCore hub.

Cover:

Core

Amazon Bedrock AgentCore
AgentCore job support
AgentCore production support
AgentCore live project support
AgentCore architecture
AgentCore deployment
AgentCore troubleshooting

Runtime

Runtime
sessions
session isolation
deployment
frameworks
scaling
long-running agents
serverless agent workloads
production runtime

Memory

short-term memory
long-term memory
personalization
conversation state
memory strategies
memory retrieval
memory troubleshooting

Gateway

tools
APIs
Lambda tools
MCP tools
tool discovery
tool authentication
tool authorization
gateway debugging

Identity

identity
workload identity
user identity
OAuth
IAM
credentials
authentication
authorization

Policy

AgentCore Policy
Cedar
least privilege
fine-grained authorization
tool policy
agent permissions

Browser

browser agents
web automation
browser tools
browser workloads
secure browser execution

Code Interpreter

sandboxed execution
Python execution
analytics workflows
code agents
file handling

Observability

CloudWatch
logs
metrics
traces
spans
sessions
tool calls
latency
failures
OpenTelemetry
production debugging

Evaluation

agent evaluation
trajectory evaluation
tool-use evaluation
response evaluation
production quality

Other current AgentCore capabilities

Research and cover where current:

Optimization
Harness
Registry
VPC/private networking
multi-agent support
A2A
MCP
GovCloud availability
enterprise deployments

Do NOT center new content on legacy Bedrock Agents if AWS now recommends AgentCore.

Preserve legacy search coverage only where useful and technically accurate.

10. AMAZON NOVA CLUSTER

Research current August 2026 Nova portfolio.

Build durable coverage around:

Amazon Nova
Nova on Bedrock
Nova text/reasoning
Nova multimodal
Nova embeddings where applicable
Nova image/video where applicable
Nova inference
Nova customization
Nova fine-tuning
Nova evaluation
Nova production
Nova prompt engineering
Nova interview support

Potential comparisons:

Nova vs Claude
Nova vs Llama
Nova vs Mistral

Do not create pages based on obsolete model-version naming.

11. AMAZON SAGEMAKER PARENT CLUSTER

Research current distinction between Amazon SageMaker and SageMaker AI.

Create a broader SageMaker parent branch covering:

Amazon SageMaker
SageMaker job support
SageMaker production support
SageMaker architecture
SageMaker implementation support

SageMaker Unified Studio

Cover:

Unified Studio
projects
notebooks
Bedrock integration
SageMaker AI integration
Redshift
Glue
Athena
EMR
collaborative development
IAM Identity Center
data/AI workflows

SageMaker Lakehouse

Cover:

SageMaker Lakehouse
S3
Redshift
Apache Iceberg
federated data
zero-ETL where current
data sharing
Bedrock + lakehouse
SageMaker AI + lakehouse

SageMaker Catalog / Data & AI Governance

Cover:

SageMaker Catalog
data/AI governance
data asset discovery
model asset discovery
metadata
lineage
data quality
access governance
Lake Formation
DataZone relationship
governed AI development

12. SAGEMAKER AI — FULL ML PLATFORM CLUSTER

Research CURRENT capabilities before creating pages.

Potential distinct pages:

SageMaker AI
SageMaker Studio
SageMaker JupyterLab
Training Jobs
distributed training
foundation-model training
fine-tuning
HyperPod
JumpStart
real-time inference
inference endpoints
serverless inference where current
asynchronous inference where current
Batch Transform
inference optimization
GPU inference
autoscaling
multi-model endpoints where current
model deployment
MLflow
Managed MLflow
Pipelines
Model Registry
Experiments
automatic model tuning
Processing Jobs
Feature Store
Model Dashboard
model evaluation
model governance
CI/CD
production troubleshooting

Do not publish pages for features that are obsolete or unavailable to new customers without marking them appropriately as legacy.

13. AWS MLOPS / LLMOPS / FMOPS

Build a strong MLOps branch.

Cover:

AWS MLOps
SageMaker MLOps
SageMaker Pipelines
MLflow
Model Registry
experiment tracking
feature management
model versioning
model promotion
CI/CD
retraining
automated deployment
rollback
artifact management
inference monitoring
model evaluation
drift where current
FMOps
LLMOps
GenAIOps

Integrate with:

CodePipeline
CodeBuild
ECR
GitHub Actions
Terraform
CloudFormation
CDK

Cross-link to existing generic MLOps pages.

14. AWS AI DATA SERVICES

Build AI-specific integration coverage around:

Amazon S3

AI training data
Bedrock data
Knowledge Bases data
ML datasets
model artifacts
vector capabilities where current

AWS Glue

Glue + SageMaker
Glue + Bedrock
ETL for ML
Data Catalog
AI data pipelines

Amazon Athena

AI data exploration
Athena + SageMaker
Athena + Bedrock
serverless analytics

Amazon EMR

EMR ML
Spark ML
PySpark ML
large-scale feature engineering
SageMaker integration
Bedrock data pipelines

Amazon Redshift

Redshift + AI
Redshift ML where current
Redshift + SageMaker
Redshift + Bedrock
enterprise analytics + GenAI

Lake Formation

AI data governance
fine-grained access
cross-account AI data

Amazon DataZone

Research current role within SageMaker governance.

Amazon MWAA

Airflow ML workflows
SageMaker orchestration
data pipelines
ML pipelines

15. AWS VECTOR / SEARCH / DATABASE AI

Build focused subcluster around:

OpenSearch

OpenSearch vector search
OpenSearch Serverless
semantic search
hybrid search
embeddings
Bedrock + OpenSearch
RAG with OpenSearch
vector indexing
production vector search

Aurora PostgreSQL

pgvector
RAG
embeddings
Bedrock + Aurora

RDS PostgreSQL

Only when distinct from Aurora intent.

DynamoDB

agent state
session state
conversation history
GenAI metadata
Bedrock application state

Neptune / Neptune Analytics

knowledge graph
GraphRAG
graph retrieval
agent knowledge

Amazon Kendra

Verify current positioning and availability before creating new pages.

16. PURPOSE-BUILT AWS AI SERVICES

Research and build current useful coverage for:

Amazon Textract

OCR
forms
tables
expenses
ID documents
intelligent document processing
Textract + Bedrock
Textract + RAG

Amazon Rekognition

image analysis
video analysis
moderation
custom labels
Rekognition + Bedrock

Amazon Transcribe

speech-to-text
streaming transcription
call transcription
Transcribe + Bedrock
voice AI

Amazon Polly

text-to-speech
neural TTS
voice AI
Polly + Bedrock

Amazon Translate

translation
multilingual applications
Translate + Bedrock

Amazon Lex

conversational AI
chatbot architecture
Lex + Bedrock
Lex + Lambda
contact-center AI

Amazon Comprehend

NLP
entities
classification
sentiment
PII
Comprehend + Bedrock

Amazon Personalize

recommendations
ecommerce personalization
real-time recommendations

Research any additional current AWS purpose-built AI services.

17. AWS HEALTHCARE AI

Research current services and capabilities such as:

AWS HealthLake
AWS HealthScribe
Comprehend Medical
Bedrock healthcare
FHIR + Bedrock
clinical RAG
PHI-safe architecture
HIPAA architecture
healthcare document AI
healthcare agents
healthcare voice AI

Cross-link existing healthcare AI/ML cluster.

Avoid duplicate generic healthcare pages.

18. AWS AI APPLICATION ARCHITECTURE

Build high-intent integration pages where appropriate:

Lambda

Lambda + Bedrock
Lambda GenAI
Lambda + AgentCore
Lambda RAG
agent tools

API Gateway

Bedrock APIs
GenAI API
authentication
rate limiting
streaming

Step Functions

AI workflow orchestration
Bedrock workflows
document workflows
agent workflows

EventBridge

event-driven AI
asynchronous GenAI
agents

SQS

async inference
queue processing
retries
DLQ
document processing

SNS

Where relevant.

AppSync

Where relevant.

DynamoDB

For session/agent/application state.

19. AWS AGENT FRAMEWORKS / STANDARDS

Research current framework support.

Cover where distinct:

Strands Agents
Strands + Bedrock
Strands + AgentCore
LangChain + Bedrock
LangChain + AgentCore
LangGraph + Bedrock
LangGraph + AgentCore
LlamaIndex + Bedrock
CrewAI + AgentCore
custom Python AWS agents

Standards:

MCP on AWS
AgentCore MCP
MCP Gateway
MCP server on AWS
authenticated MCP
enterprise MCP
A2A
multi-agent AWS

Connect to existing LangChain/LangGraph/MCP pages.

20. AWS AI COMPUTE / INFRASTRUCTURE

Research current hardware generations.

Cover durable intent around:

AWS Trainium

training
distributed training
LLM training
HyperPod
current Trainium generations

AWS Inferentia

inference
model serving
inference acceleration

AWS Neuron

Neuron SDK
model compilation
PyTorch
Trainium
Inferentia

EC2 GPU

GPU AI workloads
LLM training
AI inference
distributed training
current instance families

Do not create unnecessary pages for every transient instance SKU.

Supporting infrastructure

EFA
FSx for Lustre
S3
EBS
high-performance networking
high-performance storage

21. EKS / ECS / CONTAINER AI

Create meaningful AI-specific coverage around:

EKS

EKS AI workloads
EKS ML
EKS GenAI
EKS GPU
EKS inference
vLLM on EKS
Ray on EKS where appropriate
KServe where appropriate
Bedrock application + EKS
SageMaker + EKS architecture

ECS

ECS Bedrock applications
containerized GenAI
container inference

Fargate

serverless AI containers

ECR

model-serving images
ML container registry

Cross-link Kubernetes/DevOps clusters.

22. AWS AI SECURITY / GOVERNANCE

Build a deep enterprise branch.

Cover:

AWS AI security
Bedrock security
AgentCore security
SageMaker security

IAM

roles
policies
least privilege
Bedrock IAM
SageMaker IAM
AgentCore IAM
cross-account AI

IAM Identity Center

Unified Studio
workforce access

KMS

encryption
model/data encryption
Bedrock
SageMaker

Secrets Manager

API keys
agent tool credentials
external integrations

VPC / PrivateLink

private Bedrock
private SageMaker
VPC endpoints
enterprise networking

CloudTrail

audit
API events
AI governance

AWS Organizations

multi-account AI
SCPs

Control Tower

AI landing zones

Security Hub / GuardDuty

Only where relevant.

AWS WAF

protecting GenAI APIs
protecting chatbot endpoints

23. AWS AI OBSERVABILITY / FINOPS

Build relevant content around:

CloudWatch

Bedrock metrics/logs
AgentCore Observability
SageMaker metrics/logs
AI application monitoring
latency
token usage
throughput
failures
traces

OpenTelemetry

Especially AgentCore and distributed GenAI.

CloudTrail

Audit.

X-Ray

Where relevant.

Cost Explorer

Bedrock cost
SageMaker cost
GenAI FinOps

AWS Budgets

AI spend controls

Production cost optimization

token cost
endpoint cost
GPU cost
provisioned vs on-demand
idle infrastructure

24. AWS AI IaC / DEVOPS

Create integration coverage around:

Terraform Bedrock
Terraform SageMaker
Terraform AWS AI
AgentCore Terraform where supported
AWS CDK Bedrock
AWS CDK SageMaker
CloudFormation Bedrock
CloudFormation SageMaker
CodePipeline AI/ML
CodeBuild AI/ML
GitHub Actions AWS AI
ECR AI
CI/CD GenAI
CI/CD MLOps

Cross-link existing Terraform/DevOps pages.

25. PROGRAMMING LANGUAGE PAGES

Evaluate real intent before creating.

Potential:

Python Amazon Bedrock
boto3 Bedrock
Python AgentCore
Python SageMaker
Java Bedrock
AWS SDK Java Bedrock
.NET Bedrock
AWS SDK .NET Bedrock
Node.js Bedrock
JavaScript Bedrock
TypeScript Bedrock
Go Bedrock
Python Strands
Python SageMaker ML

Cross-link existing language clusters.

26. PRODUCTION / TROUBLESHOOTING CLUSTER

This is commercially important.

Create standalone pages where intent warrants them.

Possible Bedrock issues:

AccessDenied
IAM failures
model access errors
throttling
quotas
timeout
latency
streaming errors
SDK errors
token-limit issues
high cost
cross-region inference issues
inference-profile issues

Knowledge Base / RAG:

ingestion failure
synchronization failure
chunking problems
poor retrieval
wrong citations
metadata-filter failures
reranking problems
OpenSearch index issues
embeddings problems
hallucinations

AgentCore:

Runtime failures
Memory failures
Gateway failures
Identity/authentication problems
Policy authorization failures
Browser issues
Code Interpreter issues
MCP issues
tool-call failures
trace/observability issues
agent latency
scaling problems

SageMaker:

endpoint deployment failure
training job failure
IAM issues
container failure
GPU capacity
autoscaling
timeout
inference latency
model artifact errors
MLflow issues
pipeline failures

Each troubleshooting page should contain:

symptoms
likely root causes
diagnostic steps
logs
metrics
AWS APIs
IAM checks
configuration checks
fix
validation
prevention

No generic filler.

27. ROLE-BASED AWS AI/ML CLUSTER

Evaluate pages for roles such as:

AWS AI Engineer
AWS ML Engineer
AWS Generative AI Engineer
Amazon Bedrock Engineer
Bedrock Developer
AWS Agentic AI Engineer
AgentCore Engineer
SageMaker Engineer
AWS MLOps Engineer
AWS Data Scientist
AWS AI Solutions Architect
AWS GenAI Solutions Architect
AWS AI Platform Engineer
AWS ML Platform Engineer
AWS RAG Engineer
AWS LLM Engineer
AWS AI DevOps Engineer
AWS AI Security Engineer
AWS AI Data Engineer
AWS AI Architect

For each high-value role evaluate:

job support
interview support
profile positioning
candidate marketing
job search
interview scheduling

Do NOT automatically produce every role × every service permutation.

28. INTERVIEW SUPPORT CLUSTER

Audit and reuse existing AWS/AI interview URLs.

Create distinct pages only where intent is unique.

Potential coverage:

AWS AI/ML interview support
AWS Bedrock interview support
Bedrock technical interview
Bedrock architecture interview
Bedrock system design
Bedrock production interview
Bedrock RAG interview
Knowledge Bases interview
Guardrails interview
AgentCore interview
AgentCore system design
SageMaker interview
SageMaker AI interview
SageMaker MLOps interview
AWS ML Engineer interview
AWS GenAI Engineer interview
AWS AI Architect interview
AWS Solutions Architect AI/ML interview
AWS RAG Engineer interview
AWS Agentic AI Engineer interview
AWS AI security interview
Nova interview

Connect relevant pages to existing real interview-question pages.

Do not fabricate employer-specific interviews.

29. JOB SUPPORT / PROJECT SUPPORT CLUSTER

Create or connect intent around:

AWS AI/ML job support
AWS GenAI job support
Bedrock job support
Bedrock production support
Bedrock project onboarding
SageMaker job support
SageMaker production support
SageMaker MLOps support
AgentCore job support
AgentCore production support
AWS AI architecture support
AWS RAG support
AWS AI production troubleshooting
AWS ML project support

Use existing site terminology and funnel structure.

30. PROFILE POSITIONING / CANDIDATE MARKETING

Evaluate:

AWS AI Engineer profile positioning
AWS GenAI Engineer profile positioning
Bedrock Engineer profile positioning
AWS AI Solutions Architect profile positioning
SageMaker Engineer profile positioning
AWS MLOps Engineer profile positioning
AWS AI candidate marketing
AWS GenAI candidate marketing
Bedrock candidate marketing
AWS ML candidate marketing

Connect to existing generic AI/ML candidate marketing rather than duplicate weak intent.

31. GET INTERVIEW SCHEDULED

Evaluate meaningful pages for:

get AWS AI interview scheduled
get AWS ML interview scheduled
get AWS GenAI interview scheduled
get Amazon Bedrock interview scheduled
get SageMaker interview scheduled
get AWS AI Architect interview scheduled
get AWS ML Engineer interview scheduled

Use existing site's established architecture.

32. COUNTRY CLUSTER — REQUIRED

Build complete AWS AI/ML geographic coverage using the site's existing location model.

Audit which countries already exist across AI/ML/Angular/Workday/etc.

At minimum evaluate:

USA
Canada
UK
Ireland
Germany
Netherlands
France
Sweden
Switzerland
Australia
New Zealand
Singapore
Hong Kong
UAE
Saudi Arabia
Europe

Also inspect whether other countries are already supported by existing site architecture.

Primary country page

Each significant market should have a strong AWS AI/ML country hub where missing.

Conceptual examples only:

USA AWS AI/ML Job Support
Canada AWS AI/ML Job Support
UK AWS AI/ML Job Support

Do not blindly use these exact slug formats.

Use existing site conventions.

Country hub content should naturally cover:

Amazon Bedrock
SageMaker
AgentCore
GenAI
RAG
MLOps
major purpose-built AI services
relevant role demand
interview support
production support

Tier 1 countries

Evaluate stronger dedicated service-country pages for:

USA
Canada
UK
Australia
Germany
Ireland
Singapore
UAE

Potential dedicated pages:

Amazon Bedrock country
SageMaker country
AWS GenAI country
AWS MLOps country
AWS Agentic AI / AgentCore country
AWS AI interview country

Only create where intent/content supports it.

Tier 2 countries

Use strong AWS AI/ML parent pages containing multiple services rather than thin permutations.

33. COUNTRY LOCALIZATION

Each country page must contain real differentiated local context.

Where relevant include:

tech hubs
cloud adoption
AWS market context
major local industries
finance
healthcare
telecom
retail
SaaS
government/public sector
local job titles
local hiring language
timezone
remote/hybrid context
regional compliance considerations

Do not fabricate:

offices
employees
customers
local physical presence

34. CITY CLUSTER — REQUIRED

Follow the proven city architecture already used across the website.

Evaluate these technology hubs and additional existing cities.

USA

New York
San Francisco
San Jose
Seattle
Boston
Dallas
Austin
Houston
Chicago
Atlanta
Washington DC
Jersey City
Charlotte
Raleigh
Phoenix
Los Angeles
Tampa
Denver
Minneapolis
Columbus
Nashville
Pittsburgh
Salt Lake City

Canada

Toronto
Vancouver
Montreal
Calgary
Ottawa
Waterloo
Mississauga
Brampton
Edmonton
Halifax

UK

London
Manchester
Birmingham
Leeds
Glasgow
Edinburgh
Bristol
Cambridge
Reading
Nottingham

Ireland

Dublin
Cork
Galway
Limerick

Germany

Berlin
Munich
Frankfurt
Hamburg
Cologne
Düsseldorf
Stuttgart

Netherlands

Amsterdam
Rotterdam
Utrecht
The Hague
Eindhoven

France

Paris
Lyon
Toulouse
Marseille
Lille

Sweden

Stockholm
Gothenburg
Malmö

Switzerland

Zurich
Geneva
Basel
Bern

Australia

Sydney
Melbourne
Brisbane
Perth
Adelaide
Canberra

New Zealand

Auckland
Wellington
Christchurch

Gulf

Dubai
Abu Dhabi
Riyadh
Jeddah

APAC

Singapore
Hong Kong

Inspect existing clusters for additional locations.

35. CITY PAGE STRATEGY

Do NOT create:

CITY × EVERY AWS SERVICE

as a blind permutation.

Default city architecture should be a strong AWS AI/ML city page.

Example concept:

New York AWS AI/ML Job Support

which naturally covers:

Amazon Bedrock
SageMaker
AgentCore
AWS GenAI
RAG
MLOps
relevant local industries
roles
interview support
job/project support

Then create dedicated city + service pages only where:

search intent is distinct
demand is plausible
useful unique content exists
it does not cannibalize the parent city page

Examples that might qualify after analysis:

New York Amazon Bedrock support
San Francisco Amazon Bedrock support
Seattle SageMaker support
London AWS GenAI support
Toronto AWS Bedrock support

Do not assume they qualify.

36. COUNTRY → CITY → SERVICE FUNNEL

Create bidirectional relationships.

Example:

AWS AI/ML → USA AWS AI/ML → New York AWS AI/ML

And:

Amazon Bedrock → USA AWS AI/ML → New York AWS AI/ML

And where justified:

Amazon Bedrock → USA Amazon Bedrock → New York AWS AI/ML

Every city page should link to:

country
AWS AI/ML hub
Amazon Bedrock
SageMaker
AgentCore/GenAI
relevant interview page
relevant job support page
conversion

Every country page must link to its city pages.

No orphan city pages.

37. INDUSTRY AWS AI/ML CLUSTER

Evaluate intersections with existing industry AI content.

Potential:

healthcare AWS AI
healthcare Bedrock
financial services AWS AI
banking Bedrock
insurance Bedrock
pharma AWS AI
retail AWS AI
ecommerce AWS AI
telecom AWS AI
manufacturing AWS AI
supply-chain AWS AI
cybersecurity AWS AI
enterprise AWS AI

Each page must have technically different architecture/use cases.

Examples:

Healthcare:

HIPAA
PHI
FHIR
clinical RAG
HealthLake
Bedrock Guardrails
IAM/KMS

Finance:

PII
audit
Guardrails
agent authorization
model evaluation

Insurance:

claims
Textract
Bedrock Data Automation
RAG
agent workflows

No shallow industry templates.

38. KNOWLEDGE-BASE / GUIDE CLUSTER

Create useful evergreen informational and technical pages.

Potential:

What is Amazon Bedrock?
How Amazon Bedrock works
Amazon Bedrock architecture
Bedrock production architecture
Bedrock security guide
Bedrock Knowledge Bases guide
Bedrock RAG guide
Bedrock Guardrails guide
Bedrock inference guide
Bedrock cost optimization
Bedrock production troubleshooting
Bedrock AgentCore guide
AgentCore architecture
AgentCore Runtime guide
AgentCore Memory guide
AgentCore Gateway guide
AgentCore security
AgentCore MCP
SageMaker architecture
SageMaker AI guide
Unified Studio guide
SageMaker MLOps guide
SageMaker MLflow guide
SageMaker inference guide
AWS AI/ML architecture guide
AWS GenAI architecture guide
enterprise AWS AI architecture
how to explain a Bedrock project in interview
how to explain AgentCore in interview
how to explain SageMaker architecture in interview

Knowledge pages should link toward commercial pages naturally.

39. COMPARISON CLUSTER

Audit existing comparison URLs first.

Potential useful comparisons:

Bedrock vs SageMaker AI
Bedrock vs Azure OpenAI
Bedrock vs Vertex AI
Bedrock vs OpenAI API
AgentCore vs Agents Classic
AgentCore vs LangGraph
AgentCore vs LangChain
AgentCore vs CrewAI
Strands vs LangGraph
Knowledge Bases vs custom RAG
OpenSearch vs pgvector
RAG vs fine-tuning on AWS
Bedrock on-demand vs provisioned throughput
EKS inference vs SageMaker inference
Bedrock vs self-hosted LLM on EKS
SageMaker MLflow vs self-hosted MLflow

Do not create artificial comparisons without genuine decision intent.

40. HOMEPAGE INTEGRATION — MANDATORY

Audit homepage carefully.

Do not redesign it.

Do not remove existing high-performing content.

Add AWS AI/ML visibility into the existing appropriate section, such as:

Technologies
Trending Technologies
AI/ML
Cloud
Services

At minimum expose high-level entities such as:

AWS AI/ML
Amazon Bedrock
Amazon Bedrock AgentCore
Amazon SageMaker
AWS Generative AI

Do NOT add dozens of low-level service links to homepage.

Homepage should link to major hubs.

Major hubs should distribute authority deeper.

41. TECHNOLOGIES PAGE INTEGRATION — MANDATORY

Audit /technologies/.

Add or enhance an AWS AI/ML category.

Conceptually surface:

AWS AI/ML

Amazon Bedrock
AgentCore
Amazon Nova
SageMaker
SageMaker AI
AWS MLOps
AWS RAG
AWS Agentic AI
AWS AI Infrastructure
AWS AI Security
AWS Purpose-Built AI

Do not dump hundreds of links on /technologies/.

Use hierarchy.

42. SERVICES PAGE INTEGRATION — MANDATORY

Audit /services/.

Connect AWS AI/ML to relevant commercial services:

Services → Job Support → AWS AI/ML

Services → Production Support → AWS AI/ML

Services → Interview Support → AWS AI/ML

Services → Candidate Marketing → AWS AI/ML

Services → Profile Positioning → AWS AI roles

Services → Get Interview Scheduled → AWS AI/ML roles

Preserve all unrelated services.

43. FULL SITE FUNNEL — MANDATORY

Create a connected funnel beginning from homepage.

Conceptual architecture:

HOME ↕ TECHNOLOGIES / SERVICES ↕ AWS AI/ML ↕ BEDROCK / AGENTCORE / SAGEMAKER / AWS AI SERVICES ↕ SUBSERVICE / IMPLEMENTATION ↕ PRODUCTION / GUIDE / INTERVIEW / ROLE ↕ COUNTRY ↕ CITY ↕ CONVERSION

The graph should not be strictly linear.

Relevant pages must cross-link.

44. INTERNAL LINKING FROM EXISTING PAGES

Audit existing related pages and add contextual links from:

AI/ML
Generative AI
LLM
RAG
Agentic AI
MLOps
AWS
cloud
Python
Java
.NET
Node.js
data engineering
DevOps
SRE
Kubernetes
Terraform
LangChain
LangGraph
vector database
OpenSearch
security
responsible AI
AI Engineer
ML Engineer
GenAI Engineer
AI Architect

Do not keyword-stuff.

Only add useful contextual links.

45. NEW AWS PAGES MUST LINK BACK INTO EXISTING AUTHORITY

Examples:

Bedrock RAG → generic RAG → vector database → AI/ML

AgentCore → Agentic AI → multi-agent AI → MCP → LangGraph

SageMaker MLOps → generic MLOps → MLflow → model deployment

Bedrock security → AI security → responsible AI → AWS/cloud security

This should become a site-wide semantic graph.

46. BREADCRUMBS

Use the existing breadcrumb architecture.

Examples:

Home

Technologies AWS AI/ML Amazon Bedrock Knowledge Bases

Home

Technologies AWS AI/ML SageMaker SageMaker AI Inference

Home

Locations USA New York AWS AI/ML Job Support

Use valid BreadcrumbList structured data if already supported.

47. RELATED CONTENT MODULES

Every AWS AI/ML page should expose genuinely related technologies/pages.

Example:

Bedrock RAG:

Knowledge Bases
OpenSearch
S3
Aurora pgvector
Guardrails
AgentCore
generic RAG

SageMaker inference:

SageMaker AI
ECR
EC2 GPU
Inferentia
Trainium
CloudWatch
MLflow

Generate related links semantically, not randomly.

48. SEO REQUIREMENTS

Every indexable page must have:

unique URL
unique title tag
unique meta description
canonical
one H1
logical H2/H3
unique opening section
primary keyword
secondary keywords
relevant semantic entities
useful internal links
breadcrumbs
FAQ only where useful
conversion path
useful technical information

Avoid keyword stuffing.

Natural terms may include:

AWS AI
AWS AI/ML
AWS machine learning
AWS generative AI
AWS GenAI
Amazon Bedrock
AWS Bedrock
Amazon SageMaker
SageMaker AI
Amazon Bedrock AgentCore
AWS agentic AI
AWS MLOps
AWS RAG
AWS AI Engineer

Do not mechanically repeat them.

49. TECHNICAL CONTENT QUALITY

Technical pages must include implementation-level material where relevant.

Use terminology such as:

boto3
AWS SDK
ARN
IAM role
IAM policy
trust policy
KMS
VPC
VPC endpoint
PrivateLink
security groups
CloudWatch
CloudTrail
OpenTelemetry
request ID
modelId
inferenceProfileArn
Converse API
InvokeModel
Retrieve
RetrieveAndGenerate
Knowledge Base ID
embeddings
vector index
metadata filters
chunking
reranking
latency
tokens
quotas
throttling
retries
backoff
agent session
tool invocation
MCP
A2A
container
ECR
inference endpoint
training job
pipeline execution
MLflow run
model registry
model version

Avoid generic marketing fluff.

50. CONTENT SAFETY / ACCURACY

Do not fabricate:

customer names
case studies
results
reviews
ratings
AWS certifications
AWS partnership
AWS competency
AWS Marketplace status
offices
locations
employee counts
employer outcomes
interview outcomes
fake statistics

Do not imply official AWS affiliation.

Use AWS names accurately.

Paraphrase official technical content.

Do not copy AWS documentation.

51. STRUCTURED DATA

Inspect existing schema implementation.

Use only valid page-appropriate schema such as:

WebPage
Service
TechArticle
Article
BreadcrumbList
FAQPage
ItemList
Organization where globally appropriate

Do not create fake:

AggregateRating
Review
LocalBusiness locations

unless backed by real data.

52. GEO / AI SEARCH / LLM DISCOVERY

Inspect current implementation of:

sitemap
robots.txt
llms.txt
llms-full.txt
structured data
semantic HTML
knowledge indexes
feeds

Update existing architecture rather than creating disconnected files.

Ensure important hubs and technical pages can be understood by:

Google
Bing
ChatGPT-style search
Gemini
Claude
Perplexity
other retrieval systems

No hidden crawler-only content.

53. PAGE FRESHNESS / UPDATED DATE

Where supported by the site:

Display accurate:

Updated: August 2026

or actual modification date.

Only do this where the page was genuinely reviewed.

Use accurate dateModified.

Do not fake dates for untouched pages.

54. CANNIBALIZATION CONTROL

Before creating EVERY URL ask:

Does this URL already exist?
Does another URL target the same keyword?
Does another URL target the same user intent?
Could this be a section inside an existing page?
Does this deserve its own page?
Is this current AWS technology?
Is there enough unique technical substance?
Will the page have meaningful inbound links?
Does the page have a distinct conversion path?
Is it likely to become thin programmatic content?

If intent substantially overlaps, do not create the new page.

Enhance the existing page.

55. PROGRAMMATIC SEO RULE

Use scalable data-driven generation where appropriate.

But do NOT automatically generate:

COUNTRY × CITY × SERVICE × SUBSERVICE × ROLE

permutations.

The correct approach is hierarchical.

Example:

Global AWS AI/ML → Global Bedrock → USA AWS AI/ML → New York AWS AI/ML

Then only create:

USA Bedrock or New York Bedrock

if distinct search intent and unique useful content justify it.

Maximum legitimate coverage.

Minimum thin content.

56. CLUSTER MANIFEST — REQUIRED BEFORE IMPLEMENTATION

Before creating pages, build a structured manifest for all proposed and existing relevant URLs.

Each record should contain:

URL
page title
category
AWS service
subservice
page type
country
city
primary keyword
secondary keywords
search intent
parent page
sibling group
existing/new
stale/current
canonical target
schema
CTA
inbound link sources
outbound links
priority
implementation status

This manifest becomes the source of truth.

57. SEARCH INTENT CLASSIFICATION

Every URL should have ONE dominant search intent:

transactional
commercial
technical
troubleshooting
interview
career
local
informational
comparison

Do not allow multiple pages to target identical dominant intent.

58. IMPLEMENTATION PRIORITY

P0

Must include:

AWS AI/ML master hub
Amazon Bedrock
AgentCore
SageMaker
SageMaker AI
Bedrock Knowledge Bases
Bedrock RAG
Bedrock Guardrails
Bedrock production support
AWS MLOps
AWS AI/ML interview hub
key country hubs
homepage integration
technologies integration
services integration

P1

core technical service pages
AgentCore components
SageMaker components
Nova
OpenSearch RAG
Bedrock integrations
production/troubleshooting pages
major cities

P2

purpose-built AI services
security
observability
frameworks
data services
AI infrastructure
industries

P3

long-tail guides
comparisons
second-tier cities
specialized integration pages

Implement all pages that pass quality and collision gates.

Do not force a target page count.

59. ZERO ORPHAN PAGE RULE

After implementation, crawl the site internally.

For every new page validate:

at least one inbound internal link
parent exists
parent links to page
page links back to parent
siblings are discoverable
conversion path exists
canonical is correct
sitemap contains it if indexable

Required orphan count = 0.

60. CLICK-DEPTH RULE

Important pages should be reasonably discoverable.

Priority hubs should have short paths such as:

Homepage → AWS AI/ML

Homepage → AWS AI/ML → Amazon Bedrock

Homepage → AWS AI/ML → SageMaker

Homepage → AWS AI/ML → AgentCore

Do not create navigation requiring users/search crawlers to click through 7–8 layers to reach important pages.

Deep taxonomy is acceptable.

Deep discoverability is not.

61. SITEMAP

Update existing sitemap generation.

Do not duplicate sitemap logic.

Ensure all valid indexable pages appear.

Check:

homepage
technologies
services
AWS AI/ML hub
Bedrock
AgentCore
SageMaker
subservices
countries
cities
interviews
guides
commercial pages

Avoid duplicate URL entries.

62. ROBOTS.TXT

Do not change robots rules unless required.

If changed, preserve current valid crawler behavior.

Do not accidentally block new AWS AI/ML pages.

63. LLMS.TXT / LLMS-FULL.TXT

If the repository already maintains them:

Update them with structured AWS AI/ML hierarchy.

Prioritize:

AWS AI/ML hub
Bedrock
AgentCore
SageMaker
major service pages
major technical guides
major commercial pages

Do not dump every low-priority city URL without structure.

64. NAVIGATION

Do not create an enormous mega-menu.

Expose major hubs:

AWS AI/ML
Amazon Bedrock
AgentCore
SageMaker

Then let the hub pages expose detailed service pages.

Preserve usability.

65. HOMEPAGE SEO SAFETY

Do not rewrite the homepage.

Do not change the homepage H1/title unless truly necessary.

Do not keyword-stuff AWS terms.

Add major AWS AI/ML hubs through existing design patterns/cards/technology sections.

Protect existing homepage intent and rankings.

66. INTERNAL FUNNEL EXAMPLES

Implement relationships such as:

AWS AI/ML → Bedrock → Knowledge Bases → RAG → OpenSearch → production support

AWS AI/ML → Bedrock → AgentCore → Runtime → Gateway → MCP → Observability

AWS AI/ML → SageMaker → SageMaker AI → Training → MLflow → Pipelines → Model Registry → Inference

AWS AI/ML → AWS AI Data → Glue → EMR → Athena → Redshift → SageMaker

AWS AI/ML → Country → City → Job Support → Interview Support → Candidate Marketing

AWS AI/ML → Role → Profile Positioning → Candidate Marketing → Interview Scheduling → Interview Support

Guide → Technical implementation → Commercial support

67. BUILD / TECHNICAL VALIDATION

After all implementation:

Run the repository's actual:

build
typecheck
lint
tests
route validation
sitemap generation
static generation
content validation

Also validate:

broken internal links
duplicate routes
duplicate canonical URLs
duplicate titles
duplicate H1s
duplicate meta descriptions
orphan pages
sitemap entries
breadcrumbs
schema
responsive rendering
internal link graph
country/city relationships
service/country relationships
CTA links
homepage links
technologies links
services links

Fix every issue caused by this implementation.

Do not stop at merely generating files.

68. FINAL MANUAL REVIEW

Before completion, manually inspect representative pages from every major group:

AWS AI/ML hub
Bedrock
AgentCore
SageMaker
SageMaker AI
RAG
MLOps
purpose-built AI
security
troubleshooting
interview
role
country
city
guide
comparison

Verify the page is actually useful and not just technically valid.

69. FINAL REPORT — REQUIRED

Produce a detailed implementation report.

A. August 2026 research

List:

official AWS sources researched
major August 2026 updates incorporated
Bedrock changes
AgentCore changes
SageMaker changes
Nova changes
AI service changes
renamed products
legacy products/features
deprecated features
unavailable-to-new-customer features

B. Existing pages preserved

List relevant existing pages reused.

C. Existing pages enhanced

For each:

URL
changes made
freshness updates
internal links added

D. New master hubs

List.

E. Bedrock pages

Group by:

core
inference
models
RAG
Guardrails
Data Automation
production
security
integrations

F. AgentCore pages

Group by:

Runtime
Memory
Gateway
Identity
Policy
Browser
Code Interpreter
Observability
Evaluations
MCP/A2A
production

G. SageMaker pages

Group by:

SageMaker platform
Unified Studio
Lakehouse
Catalog/governance

H. SageMaker AI pages

Group by:

training
inference
MLflow
MLOps
deployment
governance
production

I. Nova pages

List.

J. Data/RAG/vector pages

List.

K. Purpose-built AI service pages

List.

L. AI infrastructure pages

List.

M. Security/governance pages

List.

N. Observability/FinOps pages

List.

O. DevOps/IaC pages

List.

P. Troubleshooting pages

List.

Q. Interview pages

List.

R. Role/career pages

List.

S. Candidate marketing/profile pages

List.

T. Country pages

Group by country and give counts.

U. City pages

Group by country and give counts.

V. Industry pages

List.

W. Guides

List.

X. Comparisons

List.

Y. Pages intentionally NOT created

Explain:

duplicate intent
existing page already sufficient
insufficient search intent
technically obsolete
thin-content risk
better handled as a subsection

Z. Homepage integration

Explain:

section changed
links added
major AWS hubs surfaced
confirmation existing homepage intent/design preserved

AA. Technologies integration

Show hierarchy.

AB. Services integration

Show commercial funnel.

AC. Internal linking

Report:

total AWS AI/ML pages
total inbound link coverage
total outbound link coverage
orphan count
maximum important-page click depth
country → city validation
service → country validation
existing cluster → AWS cluster validation

Required:

orphan count = 0

AD. Technical SEO

Confirm:

titles
metas
canonicals
schema
breadcrumbs
sitemap
robots
llms.txt
llms-full.txt
update dates

AE. Build validation

Confirm:

build status
tests
lint
duplicate routes
broken links
duplicate titles
duplicate H1s
duplicate metas
canonical errors
sitemap errors

70. FINAL EXECUTION ORDER — FOLLOW EXACTLY

Execute this project in this order:

Audit repository.
Inventory all existing AWS/AI/ML pages.
Research official AWS product state through August 2026.
Identify recent updates, new services, renames and legacy/deprecated features.
Compare current AWS taxonomy against existing site coverage.
Build complete AWS AI/ML taxonomy.
Build search-intent map.
Build complete cluster manifest.
Run URL collision audit.
Run cannibalization analysis.
Identify pages to preserve.
Identify pages to enhance.
Identify missing pages to create.
Implement AWS AI/ML master hub.
Implement Bedrock cluster.
Implement Knowledge Bases/RAG cluster.
Implement Guardrails/Data Automation.
Implement AgentCore cluster.
Implement Nova cluster.
Implement SageMaker parent cluster.
Implement SageMaker AI cluster.
Implement MLOps cluster.
Implement AWS AI data/vector/search cluster.
Implement purpose-built AI service cluster.
Implement infrastructure/container cluster.
Implement AWS AI security/governance.
Implement observability/FinOps.
Implement AI DevOps/IaC.
Implement programming/framework integrations.
Implement production/troubleshooting pages.
Implement role pages.
Implement job/project support funnels.
Implement interview funnels.
Implement profile/candidate marketing funnels.
Implement interview-scheduling funnels.
Implement country hubs.
Implement justified country-service pages.
Implement city hubs.
Implement justified city-service pages.
Implement industry pages.
Implement guides.
Implement comparisons.
Add AWS AI/ML to homepage using existing design.
Add AWS AI/ML hierarchy to Technologies.
Add AWS AI/ML commercial paths to Services.
Add internal links from existing AI/ML/AWS/DevOps/data/security pages.
Add reverse links from new AWS pages to existing authority clusters.
Update breadcrumbs.
Update related-page modules.
Update sitemap generation.
Update llms.txt / llms-full.txt where present.
Validate robots.txt.
Crawl full internal link graph.
Fix all orphan pages.
Check duplicate intent.
Check duplicate routes.
Check metadata.
Check canonical URLs.
Check JSON-LD.
Run build.
Run tests.
Run lint.
Inspect representative rendered pages.
Fix every problem introduced.
Produce complete final report.

FINAL MISSION

The final website architecture should work like this:

HOME ↓ TECHNOLOGIES / SERVICES ↓ AWS AI/ML ↓ BEDROCK / AGENTCORE / SAGEMAKER / AWS AI SERVICES ↓ SERVICE + SUBSERVICE + IMPLEMENTATION ↓ PRODUCTION / TROUBLESHOOTING / GUIDE / INTERVIEW / ROLE ↓ COUNTRY ↓ CITY ↓ JOB SUPPORT / INTERVIEW / PROFILE / CANDIDATE MARKETING / CONTACT

And it must also cross-link laterally so that:

Bedrock connects to RAG, GenAI, Agentic AI and AWS
AgentCore connects to MCP, LangGraph, Strands and multi-agent AI
SageMaker connects to MLOps, MLflow, Data Science and model deployment
AWS data services connect to RAG, SageMaker and Bedrock
security connects to Guardrails, IAM and responsible AI
infrastructure connects to EKS, GPUs, Trainium and inference
countries connect to cities
cities connect to service hubs
guides connect to commercial pages
interviews connect to role/profile/candidate-marketing funnels

The implementation must maximize legitimate AWS AI/ML topical authority while maintaining:

technical accuracy through August 2026
strong user experience
clear conversion paths
proper SEO hierarchy
semantic internal linking
zero orphan pages
zero duplicate search intent
minimal thin content
zero breaking changes

DO NOT begin mass content generation before completing the repository audit, August 2026 AWS research, existing-URL inventory, cluster manifest and cannibalization audit.

Once those are complete, implement the entire validated AWS AI/ML cluster end-to-end without asking for additional confirmation.

FilesExpand file tree

AWS_BEDROCK_CLUSTER.md

Latest commit

History

AWS_BEDROCK_CLUSTER.md

File metadata and controls

1. ABSOLUTE SAFETY RULE — DO NOT BREAK EXISTING SITE

2. PHASE 1 — FULL REPOSITORY AUDIT BEFORE WRITING ANY CONTENT

3. PHASE 2 — AUGUST 2026 AWS FRESHNESS AUDIT

4. MASTER SITE ARCHITECTURE

5. AMAZON BEDROCK CLUSTER

Core Bedrock

Bedrock Runtime / APIs

Bedrock Inference

Foundation Models

Bedrock Customization

Prompt Layer

Bedrock Flows

6. BEDROCK KNOWLEDGE BASES / RAG CLUSTER

7. BEDROCK GUARDRAILS / RESPONSIBLE AI

8. BEDROCK DATA AUTOMATION

9. BEDROCK AGENTCORE — FULL MAJOR CLUSTER

Core

Runtime

Memory

Gateway

Identity

Policy

Browser

Code Interpreter

Observability

Evaluation

Other current AgentCore capabilities

10. AMAZON NOVA CLUSTER

11. AMAZON SAGEMAKER PARENT CLUSTER

SageMaker Unified Studio

SageMaker Lakehouse

SageMaker Catalog / Data & AI Governance

12. SAGEMAKER AI — FULL ML PLATFORM CLUSTER

13. AWS MLOPS / LLMOPS / FMOPS

14. AWS AI DATA SERVICES

Amazon S3

AWS Glue

Amazon Athena

Amazon EMR

Amazon Redshift

Lake Formation

Amazon DataZone

Amazon MWAA

15. AWS VECTOR / SEARCH / DATABASE AI

OpenSearch

Aurora PostgreSQL

RDS PostgreSQL

DynamoDB

Neptune / Neptune Analytics

Amazon Kendra

16. PURPOSE-BUILT AWS AI SERVICES

Amazon Textract

Amazon Rekognition

Amazon Transcribe

Amazon Polly

Amazon Translate

Amazon Lex

Amazon Comprehend

Amazon Personalize

17. AWS HEALTHCARE AI

18. AWS AI APPLICATION ARCHITECTURE

Lambda

API Gateway

Step Functions

EventBridge

SQS

SNS

AppSync

DynamoDB

19. AWS AGENT FRAMEWORKS / STANDARDS

20. AWS AI COMPUTE / INFRASTRUCTURE

AWS Trainium

AWS Inferentia

AWS Neuron