I'm an AI Engineer & Data Scientist with 3.5+ years of experience building production-grade AI systems β LLM-powered agents, scalable RAG pipelines, and custom transformer-based NLP models. I write my own code end-to-end, from FastAPI backends to LangGraph agent graphs to deployment.
- π Currently building β SattvikBhojnam, a full-stack cloud-kitchen platform, plus its dedicated 6-agent AI microservice
- π± Learning β advanced LangGraph patterns, tools/function-calling, and memory persistence
- π¬ Ask me about β LLM agents, RAG systems, applied NLP, FastAPI
- β‘ Off-hours β maintaining a daily DSA streak on NeetCode (Python)
- π« Reach me β email / LinkedIn above
Production platform for a premium sattvic cloud kitchen. FastAPI + PostgreSQL backend with Redis + Celery for async workflows, Next.js 14 frontend, live kitchen streaming, OTP auth, and full order + delivery management.
π sattvikbhojnam.com Β· Backend Β· Frontend
Standalone LLM microservice powering the platform. 6-agent LangGraph supervisorβspecialist system (Router, Menu, Order, Schedule, Seva, General) with GPT-4o / GPT-4o-mini tiered routing, FAISS RAG, a Redis + PostgreSQL two-tier memory layer, and SSE streaming.
Multi-agent support bot with a supervisor router + 4 specialist agents, a versioned prompt registry, and MessagesPlaceholder-based conversational memory. A clean demonstration of production prompt management with LangChain.
π github.com/Nitesh-lng/PromptManagement-Chatbot
- LangGraph AI Agent β Multi-LLM agent with tool use & live web search.
LangGraph Β· Groq Β· OpenAI Β· Tavilyβ repo - RAG-PDF-DeepSeek β Document-based RAG pipeline.
LangChain Β· FAISS Β· DeepSeekβ repo - Fine_Tuning_PyTorch β Fine-tuning experiments in pure PyTorch (DistilBERT, LSTM) + a transformer built from scratch.
PyTorch Β· HuggingFaceβ repo
- βοΈ Email: nnitesh0101@gmail.com
- π LinkedIn: linkedin.com/in/kumarnitesh97
- π Kaggle: kaggle.com/niteshk97
Building AI agents that think, reason, and deliver at scale.

