Talk Python To Me
โ ๏ธ Any content within the episode information, snip blocks might be updated or overwritten by Snipd in a future sync. Add your edits or additional notes outside these blocks to keep them safe.
Your snips
[04:07] From Geologist to Data Scientist
[21:20] Adopt Software Engineering Mindsets
[24:06] Break Down Notebook Code
[30:31] Refactor with Conversion Tools
[36:08] LLMs Help but Need Expertise
[40:58] Exploration vs Production Code
[43:34] Standardize and Automate ML Production
[46:10] Learn DevOps Incrementally
[50:49] Share and Refactor Code
I just finished listening to this podcast:
๐ฃ๐ผ๐ฑ๐ฐ๐ฎ๐๐ Talk Python To Me with Michael Kennedy & Catherine Nelson - ๐๐ฝ๐ถ๐๐ผ๐ฑ๐ฒ: #511: From Notebooks to Production Data Science Systems
๐๐ฎ๐๐ฒ: June 28, 2025
๐๐ฒ๐ ๐ง๐ฎ๐ธ๐ฒ๐ฎ๐๐ฎ๐
Moving from exploratory notebooks to production-ready code isn't just a technical transitionโit's a signal your data science project has succeeded and is ready to deliver real value.
๐ช๐ต๐ ๐๐ ๐ ๐ฎ๐๐๐ฒ๐ฟ๐:
In finance, the gap between analysis and implementation is where many ML initiatives die. Standardizing your production workflow with frameworks like TensorFlow Extended enables consistent model deployment, automated retraining, and proper monitoringโessential for regulatory compliance and reliable financial predictions.
๐ฅ๐ฒ๐ณ๐น๐ฒ๐ฐ๐๐ถ๐ผ๐ป ๐ง
I've seen too many teams get stuck in perpetual "notebook mode" because the production transition feels daunting. This episode reminded me that this skill gap isn't insurmountableโit's about taking incremental steps toward DevOps practices while maintaining a healthy skepticism about what your models are actually doing in production.
Follow up
To get the full insight, checkout the podcast!
#TalkPythonToMe #DataScience #MLOps