AI engineering for production systems. I work with founders and engineering teams shipping LLM-powered products: evaluations, retrieval, agent workflows, and the application-layer engineering that makes a system survive production.https://shivam.dev/en-ushttps://shivam.dev/posts/build-the-eval-harness-first/https://shivam.dev/posts/build-the-eval-harness-first/An evaluation harness is not exotic infrastructure for mature LLM applications. It is the first thing to build, before retrieval, before agent loops, before any optimisation. The minimum viable version fits in an afternoon.Thu, 30 Apr 2026 00:00:00 GMTevaluationsproductionhttps://shivam.dev/posts/rag-beyond-embedding/https://shivam.dev/posts/rag-beyond-embedding/Vector search alone is not enough for production RAG. The techniques that consistently improve retrieval quality — hybrid search, rerankers, query routing, synthetic eval data — and when each one is worth the engineering cost.Thu, 23 Apr 2026 00:00:00 GMTragretrievalproductionhttps://shivam.dev/posts/reliability-is-the-pitch/https://shivam.dev/posts/reliability-is-the-pitch/The dominant pain in production LLM applications is reliability — systems that produce trustworthy answers consistently. Cost optimisation is a downstream consequence of doing reliability well, not a competing pitch.Thu, 16 Apr 2026 00:00:00 GMTproductionpositioning