Having #DataScientists Build Infrastructure & Developing Models At The Same Time Is A Terrible Anti-Pattern We’re Addicted To.

Esp at comps that aren’t early stage -- correlated w/ a lack of technical DS leadership, poor infra design, and lack of organizational alignment.

Really shows how the difference between success & failure isn’t technology choices but good project management & strategic leadership around platforms.

#mlops #mlengineering #mlplatforms #datascience

Implementing #mlops #MLplatforms:

🪄 Teh Fantasy
- Lots of greenfield infa projects
- Training DL models from scratch
- Doing data & ML infra or data not problem

🥱💤The (Boring) Reality
- Adapting or updating existing services & tools
- Integrating with #legacycode enterprise tools
- Early-stage startups buildng MLproducts off-the-ground using existing product stack w/o separate platform

Who Am I:
👉🏻 Head of #mlops & solutions at #featureform , a #virtualfeaturestore

What I Do:
⭐️ Develop #MLplatforms & #MLOpssystems that work, no matter the business or technology constraints;
⭐️ Contribute to the #opensource MLOps ecosystem & continue to drive innovation (as well as #GCP , #AWS , #Azure);
⭐️ Create high-quality & thoughtful content that helps everyone in #productionML become more productive, collaborative, & happier.