Yashasva Bafna

At Sigmoid Analytics, I build GenAI capabilities for a Fortune 500 pharmaceutical client, including a pre-call planning system that retrieves customer insights in roughly 20 seconds with 93–95% accuracy.

I've implemented Redis-based exact-match and semantic caching to reduce latency by up to eight seconds, built multilingual NER across six languages and 10+ product features, and integrated Langfuse to evaluate more than 100 production queries for groundedness, precision, and accuracy.

My work also spans practical workflow automation: I designed an AI-driven HR screening PoC that reduces manual effort by 1–2 hours per day and built a natural-language restaurant order and inventory system using LangGraph, SQL, and Python, ranked first among eight teams.

I combine data science with financial modelling and business execution. From modelling a 30-year solar project to leading marketing for an 18,000+ attendee festival, I've translated analysis into decisions, partnerships, and measurable outcomes.