Skip to content
View srimworks's full-sized avatar

Block or report srimworks

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
srimworks/README.md

Alugunulla Vigneswara Kireeti (Vignesh)

Level 5 Principal AI Systems Architect & Sovereign Engineering Lead

Autonomous Multi-Agent Swarms (Temporal / LangGraph) • Custom LLM Post-Training (QLoRA / SFT / DPO) • Raw C++/CUDA Hardware Acceleration • Zero-Trust AppSec & KMS Walletization

Hugging Face Hub VEN-Hermes-14B-AWQ LinkedIn

Featured Open-Source Engineering Repositories

Repository Domain & Core Architecture Verified Benchmark / Impact
cuda-llm-fused-kernels Raw C++/CUDA custom kernels (fused_rmsnorm_swiglu & fused_l2_cosine_similarity) using warp-level __shfl_down_sync reductions and vectorized float4 loads. 15.87x & 6.73x Speedup over PyTorch Eager on Dual NVIDIA Tesla T4 (sm_75).
sovereign-agent-skills Institutional Model Context Protocol (MCP) & Agentic Skill Library: NAWG Anti-Hallucination, 3-Pass Zero-Variance Reconciliation, and 4 Ironclad Locks Git-Ratcheted QA Loops. 100.0% Zero-Variance Convergence & deterministic self-healing QA.
zero-trust-llm-wallet-proxy Enterprise Dynamic LLM Proxy & Cryptographic Key Walletization with AES-256-GCM envelope encryption and 15-minute ephemeral in-memory leasing. Zero Static .env Exposure & 70%+ Compute Cost Reduction.
qlora-chatml-dpo-pipeline Open-weight LLM post-training harness: Rank-32 NF4 QLoRA SFT with <think> traces, Bradley-Terry DPO, and 4-bit Group-128 AWQ quantization. 99.6% SFT Loss Reduction & 100% DPO Preference Accuracy.

Flagship Production Platforms Architected

Popular repositories Loading

  1. cuda-llm-fused-kernels cuda-llm-fused-kernels Public

    High-Performance Custom C++/CUDA Kernels for LLM Serving & Vector Retrieval (15.87x Fused RMSNorm+SwiGLU & 6.73x Fused L2 Cosine Similarity via Warp Shuffles & Vectorized float4 Loads)

    Cuda 1

  2. git_docs git_docs Public

  3. srimworks srimworks Public

    Alugunulla Vigneswara Kireeti (Vignesh) — Level 5 Principal AI Systems Architect & Sovereign Engineering Lead

    JavaScript

  4. sovereign-agent-skills sovereign-agent-skills Public

    Institutional-Grade Autonomous AI Engineering Skills & MCP Governance Protocols: Epistemic Anti-Hallucination (NAWG), 3-Pass Zero-Variance Reconciliation, Git-Ratcheted Self-Healing QA Loops, and H…

  5. zero-trust-llm-wallet-proxy zero-trust-llm-wallet-proxy Public

    Enterprise Dynamic LLM Proxy & Cryptographic Key Walletization Reference Architecture: AES-256-GCM Envelope KMS, 15-Minute Ephemeral In-Memory Credential Leasing, and Multi-Cloud Compute Arbitrage

    Python

  6. qlora-chatml-dpo-pipeline qlora-chatml-dpo-pipeline Public

    End-to-End Open-Weight LLM Post-Training & Alignment Pipeline: 4-Bit NF4 QLoRA SFT with ChatML <think> Traces, Bradley-Terry DPO Alignment, and Group-128 AWQ vLLM Serving

    Python