I’m Kundan Sai — an AI Engineer working across agentic AI, RAG, knowledge graphs and backend engineering. I like turning ambiguous problems into production systems with measurable impact.
Production-oriented AI work spanning research intelligence, software development, legacy modernization and human-in-the-loop decision support.
A citation-backed research assistant over thousands of financial reports. Built the RAG foundation and evolved it into a LangGraph multi-agent architecture with specialized agents for general chat, summaries, report discovery and equity terminology.
An agentic platform for greenfield code generation and brownfield updates to existing codebases, designed to compress long software delivery cycles.
A seven-stage migration pipeline that parses legacy exports, builds a Neo4j representation of application structure and generates semantically connected .NET + Angular code.
A human-in-the-loop evaluation system combining structured extraction, rules and LLM reasoning with explainability, audit logging and PII-aware design.
I work in a lean engineering environment where developers participate in brainstorming, architecture and implementation rather than receiving predefined specifications.
My production experience has taught me how to build with modern AI frameworks. The next step is deeper first-principles understanding of machine learning, NLP, computer vision, reinforcement learning and the systems behind frontier AI.
LLMs · RAG · LangChain · LangGraph · Prompt Engineering · Knowledge Graphs · Vector Search
Python · Java · JavaScript · TypeScript · SQL · FastAPI · Spring Boot
Azure OpenAI · Azure AI Search · Document Intelligence · AWS · Docker · Jenkins · Git
Neo4j · Vector Databases · SQL Server · Embeddings · Semantic Retrieval
B.Tech — Computer Science & Engineering
AWS Certified Cloud Practitioner · Microsoft Certified: Azure Administrator Associate (AZ-104)