Skip to main content

Jun 1, 2022

Abnormal Engineering Stories, Episode #9: Productionizing Machine Learning

In episode 9 of Abnormal Engineering Stories, Dan sits down with Mukund Narasimhan to discuss his perspective on productionizing machine learning.

Key Insights

Mukund Narasimhan's ML experience spans forecasting, recommendations, and trust/safety at Airbnb, Pinterest, Facebook, Google, Amazon, and Microsoft.

ML as a serious engineering discipline has evolved significantly over the past decade, according to Narasimhan.

Unsolved ML problems remain a barrier to adoption across new industries beyond established tech platforms.

Productionizing ML—moving models from research to live systems—is a core engineering challenge discussed in this episode.

As an engineering leader at many of the world's most important tech companies, Mukund Narasimhan understands the ebb and flow of technological trends.

Computation and information have become increasingly embedded into society. In turn, the volume of data that we produce, and the computational power of the machines we use, have increased as well. Machine learning algorithms now play a role in almost every aspect of everyday life. Mukund's experience spans many of these applications, from forecasting to recommendations to trust/safety.

In this episode, Mukund draws upon his experiences at companies like Airbnb, Pinterest, Facebook, Google, Amazon, and Microsoft to share his perspective on productionizing machine learning. Mukund and I discuss:

  • How machine learning as a serious engineering discipline has evolved over the past decade

  • The biggest problems that machine learning needs to solve to reach new industries

  • The kinds of changes we can expect to see in the future

We hope you enjoy it! Don't forget to subscribe on Apple, Spotify, or Google Podcasts.

Protect Against Evolving Email Threats

See how behavioral AI detects attacks that legacy defenses miss.