Industrial AI Podcast Autonomous CNC programming!?
Apr 1, 2026
Tanmay Agrawal, Founder and CEO of Lambda Function, builds AI for precision machining. He discusses autonomous CNC programming, CAD-to-G-code recommendation engines, and closing the loop with real-time shop-floor feedback. Conversations cover training with synthetic and real data, multimodal models, sensor fidelity, and the roadblocks to full autonomy in manufacturing.
AI Snips
Chapters
Transcript
Episode notes
Company Name Came From Reinforcement Learning
- Tanmay named Lambda Function from deep reinforcement learning concepts where two-pad generation resembles a traveling salesman problem.
- The name reflects the company's aim to apply RL-inspired optimization to machining workflows.
Retrain Offline And Keep Models Private
- Use offline batch retraining and version control to avoid models drifting while incorporating shop floor feedback.
- Deliver private models for enterprise customers so historical data can accelerate learning without exposing sensitive IP.
Digital Signatures From Real Shop Floor Runs
- Closing the loop uses real-time machine process data to create a digital signature of the actual program run on a specific machine.
- That signature bridges simulation predictions from CAM with real shop-floor outcomes to improve recommendations.