Podknife - Shreya Shankar: Lessons learned after a year of putting ML into production by Machine Learning Engineered

Help
Help

Suggested Topics
- What is Podknife?
  
  Podknife is a curated podcast information and review site designed to be accessible in your browser from any device. We’re working to build the most useful podcast information source available by providing you with as much publicly available information about each podcast in our database as we can find and keeping it as up to date as possible.
- Can I listen to podcasts on Podknife?
  
  Yeah! There's a player at the bottom of the page.
- How can I get my podcast on Podknife?
  
  We’d love to hear about your podcast and consider it for addition. You can suggest your own by logging in and selecting the “Suggest a Podcast” link from the menu. Fill out the form and we’ll take it from there.
- How can I submit a review of a podcast?
  
  To submit a review of a podcast, go to the page of the podcast you want to review and click the "Submit Review" button below the image associated with the podcast. Clicking that button brings you to a new page where you can select a rating between 1 and 5 stars and write a brief review.
- How can I submit a review of a podcast episode?
  
  To submit a review of a podcast episode, go to the page of the podcast whose episode you want to review and find the specific episode you're looking for. Clicking on any episode reveals a "View Episode" button and clicking on that brings you to an episode-specific page where you can find more information displayed about the episode along with a "Submit Review" button. Clicking that brings you to a new page where you can select a rating between 1 and 5 stars and write a brief review of the episode.
- Is there a Podknife app?
  
  Not yet - we're keeping it in the browser at least for now. No app download required to access anything on the site.
- How can I change my display name?
  
  To change your display name, log in to Podknife and select "Account Settings" in the dropdown menu. From the Account Settings page, enter your desired new display name and press the "Save" button to confirm the change.
- How can I change my password?
  
  To change your password, log in to Podknife and select "Account Settings" in the dropdown menu. From the Account Settings page, enter your desired new password in the two fields below "Update Password" and press the "Save" button to confirm the change.
- How can I keep track of all my reviews?
  
  Within the menu, click on "My Profile", you will then be able to check all your reviews for both podcast & episode. You can also check all your favorite podcasts.
- How can I keep track of all my favorited podcasts?
  
  You can find podcasts you’ve favorited displayed on the My Profile page associated with your account (https://podknife.com/profile).
  You can share your My Profile page displaying your favorite podcasts as well as any reviews you’ve written publicly with a link in the form https://podknife.com/users/[displayname] where [displayname] is replaced by your Display Name.
- How can I favorite a podcast?
  
  On each podcast page you’ll find the outline of a star above the podcast title. Clicking this star fills it in and designates the podcast as a favorite (for logged in users) or invites the user to log in or register for an account (for users who are not logged in). You can find favorited podcasts displayed on the My Profile page associated with your account (https://podknife.com/profile).
- How can I register for a Podknife account?
  
  To register, click on the “Register” button in the top right of your browser window (or visit https://podknife.com/users/sign_up). Enter a Display Name to be associated with your account on the site (alongside any reviews you contribute, for example), an Email address to associate with the account and your password (twice for confirmation). Press the “Confirm” button and you should be all set, logged in and ready to review and favorite podcasts as you wish.
- How can I get in touch?
  
  We’d love to hear from you. You can reach us using the Feedback form (https://podknife.com/feedbacks/new) which can be found either in the footer on each page (for users who are not logged in) or in the dropdown menu (for logged in users).
  
  You can also reach us on Twitter by tweeting us @podknife (https://twitter.com/podknife) or on Facebook by messaging us from our page (https://www.facebook.com/podknife/).
- Log In
- Register
- Feedback
- Help
- Privacy Policy
- Terms of Use
© Podknife 2024

Please login or sign up to post and edit reviews.

This episode currently has no reviews.

Shreya Shankar: Lessons learned after a year of putting ML into production

Podcast |: Machine Learning Engineered
Publisher |: Charlie You
Media Type |: audio
Podknife tags |: Artificial Intelligence,; Business,; Careers,; Data Science,; Interview,; Machine Learning,; Technology
Categories Via RSS |: Business,; Careers,; Science,; Technology
Publication Date |: Oct 20, 2020
Episode Duration |: 01:24:00

Shreya Shankar is a Machine Learning Engineer at Viaduct AI. She's a master's student at Stanford and has previously worked at Facebook and Google Brain. She writes some truly excellent articles about machine learning on her personal blog, shankar.com/">https://www.shreya-shankar.com/

Want to level-up your skills in machine learning and software engineering? Join the ML Engineered Newsletter: http://bit.ly/mle-newsletter

Comments? Questions? Submit them here: http://bit.ly/mle-survey

Follow Charlie on Twitter: https://twitter.com/CharlieYouAI

Take the Giving What We Can Pledge: https://www.givingwhatwecan.org/

Subscribe to ML Engineered: https://mlengineered.com/listen

Timestamps:

01:30 Follow Charlie on Twitter (http://twitter.com/charlieyouai)

02:40 How Shreya got started in CS

06:00 Choosing to concentrate in systems in undergrad (shankar.com/systems/">https://www.shreya-shankar.com/systems/)

12:25 Research at Google Brain on fooling humans with adversarial examples (adversarial-examples-that-fool-both-computer-vision-and-time-limited-humans.pdf">http://papers.nips.cc/paper/7647-adversarial-examples-that-fool-both-computer-vision-and-time-limited-humans.pdf)

18:00 Deciding to go into industry instead of pursuing a PhD (shankar.com/new-grad-advice/">https://www.shreya-shankar.com/new-grad-advice/)

19:35 Why is putting ML into production so hard? (shankar.com/making-ml-work/">https://www.shreya-shankar.com/making-ml-work/)

25:00 Best of the research graveyard

29:05 Checklist for building an ML model for production

34:10 Ensuring reproducibility

39:25 Back to the checklist

44:25 PM for ML engineering

48:50 Monitoring ML deployments

53:50 Fighting ML bias

58:45 Feature engineering best practices

01:02:30 Remote collaboration on data science projects

01:07:45 AI Saviorism (shankar.com/ai-saviorism/">https://www.shreya-shankar.com/ai-saviorism/)

01:17:40 Rapid Fire Questions

Links:

shankar.com/systems/">Why you should major in systems

adversarial-examples-that-fool-both-computer-vision-and-time-limited-humans.pdf">Adversarial Examples that Fool Both Computer Vision and Time-Limited Humans

shankar.com/new-grad-advice/">Choosing between a PhD and industry for new computer science graduates

shankar.com/making-ml-work/">Reflecting on a year of making machine learning actually useful

shankar.com/ai-saviorism/">Get rid of AI Saviorism

Shreya discusses what she's learned in the past year about making ML useful by putting into production. She touches on strategies for ensuring reproducibility, feature engineering best practices, and her checklist for building AI-driven systems.