AI Society | Podcast on Programming, Coding, Machine Learning and AI

Sameer Nigam's podcast sits at the intersection of working programmer and AI practitioner, which is a spot not enough shows manage to cover well. He talks to engineers about the code they actually write, the libraries they actually hate, and the machine learning systems they actually shipped to production. The tone is casual. It feels more like a conversation at a meetup than a formal interview, and that works in its favor because guests let their guard down and admit things they would not say on a more polished show. You'll hear engineers explain why they rolled back a deployment, why one vendor's API is secretly worse than another, and what it's like trying to debug a model at 2 a.m. when stakeholders are emailing. Nigam is a good listener and doesn't try to steer conversations toward his own talking points. Some episodes are beginner-friendly walkthroughs of a single concept; others assume you know your way around PyTorch. That mix isn't for everyone, but if you're a developer trying to bridge the gap from regular software engineering into ML work, it's a useful companion. Honest and unhurried.