The Cloudcast

Building Machine Learning Apps

7 snips
Sep 21, 2022
Ask episode
AI Snips
Chapters
Transcript
Episode notes
ANECDOTE

From Banking To Building ML Startups

  • Tuhin described moving from investment banking into ML after joining a research club and working on medical prediction models.
  • He later built companies applying ML to content moderation and operational data before founding Baseten in 2019.
INSIGHT

The Real Barrier Is Production Work

  • Teams underestimate the work needed to move a model from research to a product that actually delivers business value.
  • The gap is not model quality but the engineering and integration work required to productionize models.
INSIGHT

Models Need Context And Glue Code

  • A trained model in a notebook is not directly usable because inputs/outputs are often uninterpretable and require business logic.
  • Deploying a model needs APIs, preprocessing, normalization, and UI integration beyond typical data scientist skills.
Get the Snipd Podcast app to discover more snips from this episode
Get the app