Over the last couple of months, we’ve had fascinating conversations with some of the most brilliant minds in AI. We’ve covered the golden age of specialized hardware, the importance of data wrangling to achieve reproducibility, and why programmatic data labeling is a vital step in machine learning. To wrap up the Determined Podcast Series, we sat down with OctoML CTO and co-founder Tian “TQ” Chen. TQ holds a PhD in machine learning and systems from the University of Washington, and recently accepted an assistant professor role at Carnegie Mellon University. TQ reviews the philosophy behind TVM, the problem set that OctoML is solving, and how to make ML models deployable on specialized hardware.