Retrieval eval harness
A benchmarking framework for RAG retrieval quality across chunking and re-ranking strategies. Cut hallucination rate in a production QA bot by measuring the retrieval step in isolation, not just end-to-end output.
Shivajith Mutteal / GenAI Engineer
Building and evaluating language model systems that hold up outside the demo.
I work on GenAI systems — agentic orchestration, MCPs, guardrails, and RAG middleware — and build smaller things on the side because I want to. This site collects both: the professional work and the lab, plus notes along the way. Most of it built to be read, forked, or clicked through, not just described.
Selected work
03Prompt-to-pipeline compiler
Turns a prompt spec into a structured multi-step LLM pipeline with built-in guardrails and typed intermediate outputs, instead of one long prompt doing five jobs at once.
Distilled triage classifier
Distilled a large model's labels into a small real-time classifier for support-ticket triage — 90% lower inference cost with accuracy within 1.5 points of the teacher model.
Live apps
02Eval Playground
Run the same prompt across models side by side and compare output, latency, and cost per call in one view.
Doc Q&A Demo
A small RAG app over a sample document set — built to make the retrieval step visible, not just the final answer.
Lab
03Smaller things I built because I wanted to, not because a job needed them.
side project
Chess opening trainer
Spaced-repetition drills for the openings I keep blundering, built after losing the same line four times in a row.
side project
Trail route generator
Give it a distance and a starting point, get back a runnable loop instead of an out-and-back.
weekend build
Recipe remix bot
A just-for-fun tool that riffs on whatever's actually in the fridge instead of sending you shopping.
Writing
03What breaks first when you scale a RAG pipelineJun 2026
Notes on evaluating LLM evalsApr 2026
A small classifier beat our 70B model — here's whyFeb 2026