At Aayu Smart Health Tech, I launched Ask Aayu, a RAG health assistant that answers users’ medical-history questions on iOS and Android. I grounded 98% of clinical values in retrieved rows and cut conversation costs 40% while maintaining the same evaluation pass rate.
I also built a prompt-evaluation harness that gates releases on a 95% pass rate, and deployed the FastAPI, Redis, and Langfuse stack to AWS EC2. Tracing each turn helps track token use, model costs, and retrieval and tool latency.
At App Orchid Inc, I launched Easy Answers, an enterprise Q&A platform, on GCP in three weeks. LangChain and LangGraph agents with conditional model routing cut LLM costs 80%, while Python RAG and ReAct pipelines powered NLP and Text-to-SQL retrieval.
Earlier, as a Senior Software Engineer, I built Contract AI, which cut contract review and negotiation time 60%, and developed React and TypeScript workflows for contract redlining and term comparison. I also led three product releases and shipped the company’s first mobile product end-to-end.

