Skip to content
View TengKianBoon's full-sized avatar
💭
Building and hardening AI applications with a business-first lens
💭
Building and hardening AI applications with a business-first lens

Block or report TengKianBoon

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
TengKianBoon/README.md

Kian Boon (John) Teng

AI Solutions Architect · hands-on Python and agentic AI · C-suite business judgment · Singapore–Indonesia

Boardroom to codebase. I turn business problems into working AI solutions, from the decision and architecture through implementation, testing and release. I am Co-Founder and COO of 180Climate, building with Claude Code, Claude Cowork and Codex and owning the delivery decisions.

My edge is the combination: hands-on AI delivery with the commercial and operating judgment to identify the problem, test assumptions and connect the solution to how the business works.

Focus: AI Solutions Architect, senior forward-deployed AI delivery and climate leadership with an AI edge. Connect on LinkedIn.

Applications built around business decisions

Application What I built and why Explore
Carbon Pre-Feasibility Screening — live Qualify early development opportunities: turn investment questions into explicit assumptions, traceable calculations and reviewable screening results before further diligence and capital commitment. Try Carbon · Architecture and contribution
EUDR Plot Check — live Prioritise export due-diligence work: connect geospatial inputs, typed APIs and explicit evidence states to the plots and evidence needing review. Try EUDR · Decision logic
Fieldwork — native controlled public beta Connect project demand with field-service capability: private intake and operator review, SQLite records, consent before contact sharing and recovery procedures. Public entry · Dated verification

How I build and improve AI applications

  • Translate the problem into a working design. Map the user’s decision, business purpose and operating responsibilities to data, workflows, APIs and acceptance criteria.
  • Build hands-on with an agent team. Configure implementation, review, verification and testing roles, scoped context and bounded retries. Inspect code, run browser QA and work through release corrections.
  • Make numerical results reproducible. Python/FastAPI services, typed contracts and deterministic engines connect inputs to calculations and evidence states; models assist where interpretation is useful.
  • Build controls into the workflow. Operator access, private/public data boundaries, structured SQLite records, consent and human-review points support the operating use.
  • Keep proof close to the implementation. Architecture decisions, source, regression tests and dated delivery records make the work inspectable. The documented 1 July 2026 CI result is 461 tests passed and 2 skipped.

Hardening AI-generated code: two concrete examples

At 180Climate, I worked through corrections that connect domain judgment to application behaviour:

  • Peat routing: an independent review identified an unsuitable methodology reference. The route was corrected and golden-case tests prevent that reference returning. Decision and review context · Regression tests.
  • Uncertainty propagation: density and loss-rate variability were added to the estimate range, with the non-permanence buffer applied separately. I approved the documented change and worked through the release correction. ADR-0016.

This is how I improve AI-generated applications: locate the weak assumption, inspect the data and code path, correct the behaviour and preserve the result in tests.

Evaluation, knowledge and automation

Project Mechanisms and contribution Evidence
LLM Decision Lab Browser workspace for supplied model answers: weighted criteria, blind labels, claim comparison, local heuristic scoring, disagreement views and two-pass human review. Exported judge prompts support a separate model-assisted assessment. Project · Evaluation logic
Governed Audio Learning Pipeline Transcription, synthesis, review and export connected to private source storage, reusable tool contracts, cost/retry controls and reviewed public release. MCP-ready interfaces and runbooks make the stages inspectable. Project · Operating controls
AI Vendor Presentation Monitor Configuration-driven discovery, trusted-source filtering, deduplication, optional transcript processing and digest preparation. Project
APMA multilingual audio workflows Regional speech and long-file processing, including Qwen for Hokkien and MERaLiON for Singlish; conversion, chunking, stitching, provider comparison and speaker review. Project

The executive edge behind the delivery

  • 180Climate — Co-Founder and COO. Lead four Indonesian REDD+/IFM development opportunities: screening, commercial preparation, fundraising strategy and partner coordination. Built the commercial and financing model for phased capital, landholder/community economics, returns and offtake pathways.
  • Gemalto Indonesia — President Director and Country Manager. Led country operations, executive customer relationships and multi-vendor consortia for national eID and ePassport tenders, connecting technical and commercial decisions with government stakeholders.
  • PT Ocean Metal Indo — Chief Operating and Commercial Officer, 2011–2023. Originated and led investment, financing, development and operations for a regulated 11,260-hectare asset, from greenfield into production and a managed divestiture process. Built infrastructure, operating functions and teams; coordinated customers, regulators, communities and delivery partners.
  • Aserra Partners — Vice President of Business Development. Sustainable-infrastructure opportunity development, waste-to-energy consortium coordination and a heavy-haulage electrification pilot in Indonesia.

180Climate’s affiliated PT Global Alam Lestari (GAL) project has VCS/CCB certification and Natural Climate Solutions Lighthouse recognition; BeZero recorded an A rating in November 2025.

Business purpose, architecture and feedback

I begin with the customer or investment decision and its economics. I map responsibilities and delivery dependencies, build the workflow, then use operating feedback to decide what to improve, expand or stop. Executive experience helps me ask the right commercial questions; hands-on AI capability lets me turn them into a working solution.

Foundations: enterprise IT consulting, presales and channel work, including Oracle business-process solutioning; NTU FlexiMasters in Business AI and Technology (2026), 15 Academic Units, CGPA 4.80/5.00; Executive Mentor, Singapore Leaders Network; telecom-security patent co-inventor.

Let’s discuss the business problem and the AI solution you want delivered. LinkedIn · 180Climate

Pinned Loading

  1. 180climate-app 180climate-app Public

    Platform behind 180Climate's free carbon pre-feasibility and EUDR plot-check tools: one geospatial core, deterministic engines, typed contracts and governed AI delivery.

    Python

  2. llm-decision-lab llm-decision-lab Public

    Browser workspace for comparing supplied model answers: weighted criteria, blind labels, local scoring, disagreement views and two-pass review with human escalation.

    JavaScript

  3. governed-audio-learning-pipeline governed-audio-learning-pipeline Public

    Local-first knowledge workflow: transcription, synthesis, review and export with private source storage, cost/retry controls and approval before public release.

    HTML

  4. ai-vendor-presentation-monitor ai-vendor-presentation-monitor Public

    Configuration-driven discovery of official AI presentations: trusted-source filtering, deduplication, optional transcript processing and digest preparation.

    Python

  5. apma-singapore-asr apma-singapore-asr Public

    Evidence-preserving transcription for Hokkien, Singlish, Mandarin and English—multilingual ASR, diarization, cost controls, provenance and human review.

    Python