Building reliable Agentic RAG, multi-agent systems and production-ready LLMOps platforms.
- 6+ years building AI/ML and cloud-native systems.
- Architecting production Agentic RAG for knowledge retrieval across 5,000+ enterprise SOP documents and ServiceNow tickets.
- Building LangGraph supervisor-worker orchestration for 25+ specialized agents using MCP-enabled tools and A2A communication.
- Improving grounding with BM25 + dense hybrid retrieval, RRF, reranking, query rewriting, contextual compression and semantic caching.
- Implementing evaluation and observability with RAGAS and Opik across faithfulness, relevance, latency, token usage and agent/tool traces.
- Document Portal — multi-LLM document processing setup covering model providers, embeddings and vector database options.
- MultiPDF Bot — LLM-based assistant for querying multiple PDF documents.
- Sentiment Analysis Web App — end-to-end NLP application for sentiment classification.
- ML & Deep Learning Projects — practical machine-learning and deep-learning experiments.
Context engineering, reliable retrieval, agent orchestration, evaluation, observability and guardrails for production enterprise AI.