Yogesh J.

I build AI agents, chatbots, RAG systems and LLM products that run in production, not just in demos. 8+ years of Python backend engineering, the last 3+ focused on LLM systems used by real customers. What I can build for you: - AI agents that run real operations: multi-agent orchestration, tool calling, memory, and human approval before anything important happens - RAG over your documents and data: hybrid search (semantic + keyword), reranking, citations, and evaluation so you know the accuracy before launch - AI chatbots and voice agents for support and sales: answers from your knowledge base, human hand-off, on web, WhatsApp or phone - LLM features inside your product: plain-English Q&A, summaries, scoring, classification, multilingual output, real-time voice - AI-powered integrations: connectors between your SaaS tools that map data, recover from errors and heal themselves Production work: - CampaignSwift (founder): a multi-agent platform where AI agents plan content, chase client approvals, triage the inbox, recover failed posts and flag at-risk clients for social media agencies. ReAct agents with tool calling, human-in-the-loop, full audit log. - StockGenie AI: AI stock research for 2,000+ NSE stocks with an "Ask Me" assistant in English and Hindi. 4,000+ installs, rated 4.5 on Google Play. - Above AI: backend for a wearable AI pendant with real-time voice transcription and "chat with your own memory" (35,000+ conversations processed). How I work: clear scope first, a working demo in week one, daily updates, evaluation numbers (accuracy, latency, cost per query) before launch, and clean documented code you own. 95 jobs on Upwork with 100% Job Success and a 5.0 rating. Stack: Python, FastAPI, Django, Claude, OpenAI, Llama, LangChain, LangGraph, pgvector, Pinecone, Chroma, LangSmith, Celery, Redis, PostgreSQL, WebSockets, Docker, AWS, GitHub Actions. Send me the problem you want solved and I will reply with a short plan and a fixed quote.