Why are Machine Learning Projects so Hard to Manage?

Date: 
Monday, March 18, 2019 - 18:30
Source: 
SF Bayarea Machine Learning
Attendees: 
167
City: 
San Francisco

Main Talk: Why are Machine Learning Projects so Hard to Manage? - Lukas Biewald (Founder, Weights & Biases https://www.wandb.com/)

Abstract: I’ve watched lots of companies attempt to deploy machine learning — some succeed wildly and some fail spectacularly. One constant is that machine learning teams have a hard time setting goals and setting expectations. Why is this?

Bio: Founder of CrowdFlower, Founder of Weights and Biases

Lightning Talk: Personalizing Notification Frequency at Nextdoor
- Amit Lakhani (Engineering Lead, Nextdoor https://nextdoor.com/)

Abstract: Notifications at Nextdoor provide a lot of value to our members. Recently, we have been making inroads at personalizing the frequency that these notifications reach members, since not all members love us equally! In this short talk, I will walk through some of our current explorations as we work to optimize and personalize how many notifications our members want to receive.

Bio: Amit Lakhani got his Ph.D. at U.C. Berkeley in lasers and physics. He then worked in technical consulting, where he worked on a multitude of issues like fire investigations, electromagnetic simulations and modeling, hardware development, failure analysis, and source code forensics. He then started his own company aimed at providing lawyers with quality technical peer review, which he ran for 3 years. At Nextdoor, he is an engineering lead working on the Notifications and Content Quality and Ranking teams and is actively hiring!!!

Tentative Schedule:

6:30pm-7:00pm -- pre-reception

7:00pm-7:15pm -- lightning talk

7:15pm-8:00pm -- main talk

8:00pm-8:30pm-- post-reception

SPECIAL THANKS

Thanks to Nextdoor (https://nextdoor.com/) for hosting, snacks/drinks, and video recording!

Nextdoor HQ

875 Stevenson St #700