I am a Principal AI Engineer specializing in the intersection of Generative AI, Cybersecurity, and MLOps I focus on moving beyond simple LLM calls to building autonomous, deterministic agentic systems that solve complex enterprise challenges.
- 🤖 Currently Focused On: Architecting agentic workflows for SOC automation and proactive threat hunting.
- 🏗️ Core Expertise: RAG & GraphRAG optimization, Agentic AI (Function Calling and Tool Calling), and productionizing LLMs at scale.
- 🎙️ Community: Former Rasa Community Chapter Lead (Chennai & Bengaluru); organized 12+ meetups for 300+ members.
- ✍️ Writing: I maintain Whispering Wasps, focusing on how Markdown and reasoning models are redefining the AI stack.
- AI/ML: Generative AI, RAG, GraphRAG, Agentic Workflows, LangChain, LlamaIndex, CrewAI, Rasa.
- NLP/CV: NER, OCR (EasyOCR, Tesseract), ASR/STT (Deepgram), TTS (Cartesia).
- Backend & Infra: FastAPI, Docker, Kubernetes (GCP/Azure), MLOps (Datadog, Wiz.io, Nexus IQ, ArgoCD).
- Databases: MongoDB, SQL, Vector Databases.
- Agentic Workflows: Reviewed and designeed a 24/7 autonomous SOC detection system with response latency < 2 minutes.
- RAG Optimization: Improved retrieval accuracy by 20% and reduced latency by 60% through structured chunking strategies.
- Production Scale: Deployed conversational agents handling 100k+ monthly conversations with 99.7% uptime.
- Why Markdown Might Be All We Need: My Take
- Stop Engineering Context, Start Predicting It
- Reasoning vs. Rules: How Claude Mythos Broke Security’s Old Game
- REASONING vs RULES : Part 2 — The Feedback Engine: How the Colosseum Learns Faster Than Threats Evolve
- Productionize ML/DL Models as Easy as a Pie: BentoML
