Senior Staff AI/ML Engineer · Agentic AI in Production · Forward-Deployed Delivery
I build multi-agent LLM systems that actually run in production — and reach real users.

• Architect of the production multi-agent system behind MCX (MasterClass Executive) — an AI-native executive program built with UChicago Booth, teaching 80+ topics live to 705 executives across 55 countries.
• 4-agent architecture (orchestrator · teacher · memory · guardrail) on frontier LLMs, with plan-and-execute control, long-term memory, and an evaluation & debugging harness for accuracy, safety, and reliability.

My repos have 10K+ stars and a 2.3K-follower community:

• Deep-Learning — 🌟 3.4K — hands-on neural-network notebooks
• chat-with-code — 🌟 73 — chat with GitHub repos using LangChain + LLMs
• face_and_emotion_detection — 🌟 325 — real-time computer vision
• aerial_pedestrian_detection — 🌟 106 — drone object detection

• Agentic AI: multi-agent orchestration · ReAct & plan-and-execute · agent memory · tool/function calling · guardrails · agent evals & observability
• LLMs: RAG · fine-tuning (LLaMA family) · synthetic data · MCP · custom Claude skills
• Also: computer vision · NLP · production ML at enterprise scale (incl. regulated finance)

• 📝 1M+ views on my applied-AI writing — priya-dwivedi.medium.com
• 🎙️ Speaker, Google NEXT 2018 (production deep learning with TensorFlow)
• 📬 New Substack on shipping agents in production — coming soon

M.S. Computer Science, Georgia Tech (CV & Robotics, GPA 4.0) · Graduate Certificate in AI, Stanford · B.Tech, IIT Delhi. Previously founded and ran an applied-AI practice (100+ enterprise systems) and led ML in regulated financial services.

Beyond the demo — building AI agents that survive contact with production.

Beyond the demo — building AI agents that survive contact with production.