Manas Sisodia

I build and ship production-grade AI systems — agentic workflows, RAG pipelines, and the event-driven backends that power them. Not demos: my work runs live in production, including a generative-AI chatbot built for a national government sports program during my current role as an Agentic & Generative AI Engineer.

What I can help you build:

Agentic AI — multi-step, tool-using LLM workflows (LangChain, LangGraph)
RAG systems — end-to-end ingestion to retrieval, with reranking, HyDE query expansion, and vector search (pgvector, Pinecone, Voyage AI, Cohere)
LLM integration — OpenAI & Anthropic APIs, streaming, structured outputs
Scalable backends — Node.js, FastAPI, WebSockets, Redis; one of my real-time pipelines sustains 300+ events/sec with zero event loss under peak load
Cloud & DevOps — AWS, Docker, CI/CD (I cut one team's deploy time by 80%)

Recent production work: a public-facing GenAI chatbot for a national youth-sports program (live, ~10k daily active users); an internal, authenticated AI tool that helps government staff prepare parliamentary-question responses, in live use during an active session; and an internal task-tracking system for a national sports authority (in production).

I care about systems that hold up under load and stay maintainable — clean architecture, no event loss, code the next engineer can read. I've placed in 3 hackathons and rank in the global top ~15% on Codeforces.

I'm looking for long-term engagements where I can own your AI or backend work over time — build it, maintain it, and scale it alongside your team, rather than a one-and-done gig. If that's the kind of partner you need, message me and tell me what you're building.

Note on the first two lines: Upwork truncates the overview in search results, so the hook (production AI + live government chatbot) has to land before the "…more" cutoff. Keep those first two sentences intact.