Making LLMs hallucinate less, how to diagnose ML models, from PM in Google AI to CEO of Galileo - Vikram Chatterji - The Data Scientist Show #066

The Data Scientist Show

Vikram is the co-founder of Galileo – an AI diagnostics and explainability platform used by data science teams building NLP, LLMs and Computer Vision models across the Fortune 500 and high growth startups. Prior to Galileo, Vikram led Product Management at Google AI, where his team built models for the Fortune 2000 across retail, financial services, healthcare and contact centers. He has a master degree from Carnegie Mellon University from the school of computer science. If you enjoy the show, subscribe to the channel and leave a 5-star review. Subscribe to Daliana's newsletter on www.dalianaliu.com for more on data science and career.

Resources:LLM Studio: https://www.rungalileo.io/blog/announcing-llm-studio

Galileo: https://www.rungalileo.io/

Blog on LLM Hallucination: https://thesequence.substack.com/p/guest-post-stop-hallucinations-from

Vikram Chatterji’s LinkedIn: https://www.linkedin.com/in/vikram-chatterji/

"The Mom Test": https://www.amazon.com/The-Mom-Test-Rob-Fitzpatrick-audiobook/dp/B07RJZKZ7F

Daliana's Twitter: https://twitter.com/DalianaLiu

Daliana's LinkedIn: https://www.linkedin.com/in/dalianaliu

(00:00:00) Introduction

(00:04:24) How he got into machine learning

(00:06:53) Diagnosing large language models

(00:09:56) Addressing model hallucination

(00:12:46) Metrics for measuring hallucination

(00:17:30) From Google AI to starting Galileo

(00:24:08) Developing LLMs and putting them into production

(00:32:51) Galileo's diagnostics and explainability platform

(00:43:16) Advice for data scientists when joining a startup