Improving Code Reviews with Github’s Copilot

Paige is the director of Machine learning and machine learning operations, aka MLOps, at GitHub. Before that, she was a principal product manager at Microsoft and also worked on DeepMind and Google Brain. Paige has had over a decade of experience with machine learning and data science as a practitioner. 

Check out my new project awesomecodereview.com workshop!

Links:

• Retweet and like to win access to GitHub Codespace, including Copilot
• Tiferet’s work, using machine learning to detect security vulnerabilities in source code.
• VS Code’s Python extension and Jupyter extension.
• Copilot website (make sure to download the Copilot Nightly extension, to get the latest features!)
• Applied Machine Learning Scientist – Microsoft job opening here!
• Github – Use it for your work and tell us how we can improve!

Shownotes:

[00:01 – 10:53] Opening Segment 

• Check out my latest project: Awesome Code Reviews!
  • Visit https://www.awesomecodereviews.com/ to find articles about code reviews, best practices, code review checklist, news about the latest research and code reviews, and workshops and courses about this topic
• Get a chance to try out GitHub Codespaces and other extensions like GitHub Copilot!
  • Like and retweet today's episode, and for an additional chance to win, you can also leave a comment about what kind of data science work you're currently doing or what you like to do
• The responsibilities of a director of machine learning and machine learning operations
• Demystifying the process of reviewing complicated data science code

[10:54 – 20:54]  A Helpful Collaborator As You Write Codes

• How GitHub Copilot becomes your partner and collaborator when writing codes
  • It is an extension for VS Code and generates source code
• Learning from test cases and how code reviewers can perform a better job
• Acquiring accurate code snippets through understanding the specific requirements
• The strive for consistent performance across every single kind of language

[20:55 – 35:25]  Expanding Feature Capabilities for Optimal Functionality

• The beginning of deep learning techniques application
  • This targets detecting security vulnerabilities through code reviews
  • It also provides recommendations for extracting functions from blocks of code
• Encouraging consistency in names and styles
  • Take note: Microsoft is hiring!
• Striking the balance with deep understanding of data-driven and quantitative approaches
  • Data can tell us about users who are already using our tools, but not about those who haven't tried them yet
  • The key is to remain curious and constantly seek to better understand users

[35:26 – 37:52] Closing Segment

• Paige’s recommendation for you
  • Try out GitHub for your machine learning projects!
• Final words

Resources Mentioned: 

• Retweet and Linke this tweet to win access to GitHub codespaces and copilot
• Awesome Code Reviews - Visit for helpful information and courses for you to try!
• Applied Machine Learning Scientist - Microsoft job opening here!
• Github - Use it for your work and tell us how we can improve!

Tiferet's work, using machine learning to detect security vulnerabilities in source code.

VS Code's Python extension and Jupyter extension.

Copilot website (make sure to download the Copilot Nightly extension, to get the latest features!

Let’s Connect! You can connect with me, Dr. McKayla on Instagram, Twitter and Youtube to look into engineering software, and learn from experienced developers and thought leaders from around the world about how they develop software!

LEAVE A REVIEW + help someone who wants to know more about the engineering software world. Your ratings and reviews help get the podcast in front of new listeners.

Episode source