Engineering notes
Write-ups from real work: the design decisions, trade-offs and mistakes behind the AI systems we ship.
In VakSutra AI, code computes the score and the LLM only explains it — with verbatim quotes, real timestamps and strict rules. Here is how the split works and what we would add next.
Real-time face tracking in the browser sounds better, but it drops frames on ordinary laptops. In VakSutra AI we analyse the recording right after the user stops — here is the reasoning and the trade-offs.
Not every workflow needs an AI agent. A practical guide to choosing between fixed automation, an LLM step inside a workflow, and a full agent — and how to keep each one safe.
Why basic vector search breaks down on financial filings and complex documents, and how we architected a production-level RAG engine with intent classification, cross-encoders, and parallel agent execution.
Building something similar?
We reply within 24 hours with questions or a first outline of scope and approach.
