Gen AI Developer / Backend Engineer with 5+ years of software experience building and operating enterprise Gen AI soluti

github

ram sankar  • Peerlist

Designed and developed a generative AI-powered medical chatbot using LangChain and GPT models, integrated with Pinecone for vector-based retrieval of relevant medical contexts. Deployed a Flask-based backend with AWS CI/CD pipelines, improving response relevance and reducing end-to-end response latency by approximately 35% through optimized retrieval and prompt orchestration.

MakerBOT: AI-powered educational visualizatio

∙Developed a full-stack travel planner using TypeScript, React, and Google Gemini API for AI-powered itinerary generation with Google Maps Platform integration for route optimization ∙Built real-time expense splitting with WebSocket synchronization and Redux Toolkit state management for collaborative budget tracking across travel groups

NomadSync: AI-powered collaborative travel pl

•Contributed to Meta's PyTorch OpenEnv library for RL post-training, implementing environment wrappers, reward shaping modules, and observation preprocessing utilities ∙Applied learnings to win 1st place at Meta AI/AMD/PyTorch Synthetic Data Hackathon for LLM fine-tuning using GRPO with custom reward weights

OpenEnv (Meta PyTorch Contribution): Open-sou

∙Architected production-ready GraphRAG pipeline using LangChain, Neo4j, and GPT-4o to extract startup intelligence from TechCrunch, enabling semantic search across 1000+ articles with relationship-aware context retrieval ∙Implemented OpenAI embeddings with vector similarity matching and Cypher query interface for multi-hop graph traversals across startup ecosystems, funding patterns, and investor networks