Best Practices for Using Jupyter Notebooks in ProductionKey Takeaways:
- Discuss the pros and cons of using a notebook in a production-oriented environment.
- Explore the blind spots users have when using a notebook in production.
- Based on cross-disciplinary research done by the DagsHub team, we'll cover the best practices for using both Jupyter Notebook and IDEs that enable us to iterate faster.
Jupyter Notebooks have seen enthusiastic adoption among the data science community to become the default environment for research.
However, transitioning a project hosted in Jupyter Notebooks into a production-ready code can be a challenging task. The non-linear workflow, lack of versioning capabilities, inadequate debugging and reviewing tools, integration with development environments, and more made the productionization process an uphill battle.
Should we just throw our Jupyter Notebooks out the window? Absolutely not. After all, they are a great tool that gives us superhuman abilities. We can, however, be more production-oriented when using them.
In this session, we'll share 7 guiding principles developed over the course of 4 years of research, that help many teams and individuals scale their work, better utilize Jupyter Notebooks, and successfully bring projects from research to production.
Always pushing the envelope and exploring the frontiers of technology. Nir combines a unique background of computer vision engineering, MLOps research, and public speaking - to give a fascinating session on topics he lives and breathes.
Nir is leading the data science, MLOps, and outreach activity of DagsHub worldwide. He focuses his research on improving workflows for data science teams that work in a production-oriented environment.
Nir graduated with honors from the BGU structural engineering faculty, majored in Structural Analysis and Finite Element Simulations, and is currently pursuing his Master's in Data Science from Reichman University.