Machine Learning Street Talk (MLST)

#76 - LUKAS BIEWALD (Weights and Biases CEO)

Jun 9, 2022

Lukas Biewald, the CEO of Weights and Biases, shares his insights as a successful entrepreneur in the AI space. He discusses the recent $15 million funding round for his company and the challenges of improving training data quality. The conversation touches on the balance between generalization and specialization in machine learning, alongside the critical need for explainability in AI tools. Biewald also highlights innovative community engagement strategies through YouTube sponsorships, emphasizing the evolving nature of machine learning and its entrepreneurial dynamics.

57:36

Creator website

Episode guests

Lukas Biewald

AI Chapters

Episode notes

Intro

2min

Funding Insights and Entrepreneurial Journey

2min

Navigating Machine Learning in Gaming and Beyond

20min

Navigating the Challenges of Generalization and Specialization in ML DevOps

4min

Navigating Hyperparameter Optimization and Model Explainability

18min

Community Engagement and Algorithm Strategies in Machine Learning

4min

Evolving Machine Learning and Entrepreneurial Dynamics

8min

Check out Weights and Biases here!

https://wandb.me/MLST

Lukas Biewald is an entrepreneur living in San Francisco. He was the founder and CEO of Figure Eight an Internet company that collects training data for machine learning. In 2018, he founded Weights and Biases, a company that creates developer tools for machine learning. Recently WandB got a cash injection of 15 million dollars in its second funding round.

Lukas has a bachelors and masters in mathematics and computer science respectively from Stanford university. He was a research student under the tutelage of the legendary Daphne Koller.

Lukas Biewald

https://twitter.com/l2k

[00:00:00] Preamble

[00:01:27] Intro to Lukas

[00:02:46] How did Lukas build 2 sucessful startups?

[00:05:49] Rebalancing games with ML

[00:08:14] Elevator pitch for WandB

[00:10:38] Science vs Engineering divide in ML DevOps

[00:14:11] Too much focus on the minutiae?

[00:18:03] Vertical information sharing in large enterprises (metrics)

[00:20:37] Centralised vs Decentralised topology

[00:24:02] Generalisation vs specialisation

[00:28:59] Enhancing explainability

[00:33:14] Should we try and understand "the machine" or is testing / behaviourism enough?

[00:36:55] WandB roadmap

[00:39:06] WandB / ML Ops competitor space?

[00:44:10] How is WandB differentiated over Sagemaker / AzureML

[00:46:02] WandB Sponsorship of ML YT channels

[00:48:43] Alternatives to deep learning?

[00:53:47] How to build a business like WandB

Panel: Tim Scarfe Ph.D and Keith Duggar Ph.D

Note we didn't get paid by Weights and Biases to conduct this interview.

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.