Virtual WiMLDS x Affirm Machine Learning Talks
Details
Please join us for a virtual event, featuring women in Machine Learning@Affirm. You will learn about different topics the team at Affirm works on, ranging from building models for optimal fraud recovery, hyperparamenter tuning for imbalanced data sets, and ways of applying ML to new markets in Finance!
We’ll have a series of lightning talks followed by two small sessions of speed networking via Zoom's breakout rooms and a small survey for individuals interested in knowing about opportunities at Affirm.
Link: https://forms.gle/qCboxVfUd8FD9KP98
Everyone supporting our mission, regardless of gender, is invited to attend. Please keep in mind our Code of Conduct applies to all meetups: https://github.com/WiMLDS/starter-kit/wiki/Code-of-conduct
Zoom link will be sent out day of event
SCHEDULE
6:00 PM: Welcome and Introductions by Affirm and WiMLDS (Erin LeDell)
6:15 PM: Kelli-Jean Chun- "Predicting High LTV Users for a Facebook Marketing Campaign".
6:30 PM: Vicky Wang- "Optimal queuing for fraud recovery".
6:30 PM - 6:45 PM: Breakout session - two groups
6:45 PM: Cristine Marsh- "Who measures the metrics".
7:00 PM: Fang Fang- "Applying machine learning to new markets".
7:15 PM - 7:30 PM: Breakout session - two groups and some networking
BIOS & ABSTRACTS
Kelli-Jean Chun
Bio: Kelli-Jean a ML Scientist on Affirm’s Applied ML Personalization team and I’m currently working on user lifetime value (LTV) and recommendations. Prior to Affirm, I was a Data Scientist at Turo working on search and recommendations.
Talk: “Predicting High LTV Users for a Facebook Marketing Campaign.” Kelli-Jean will cover a binary classification model used to generate a list of recently acquired customers with high predicted LTV, which was used for a “new to Affirm” marketing campaign on Facebook.
Vicky Wang
Bio: Vicky is a machine learning scientist at Affirm, focusing on fraud prevention. She earned her Master's at UCSB before joining Affirm.
Talk: “Optimal queuing for fraud recovery.” This talk will cover how we use machine learning to select which transactions should be reviewed by our agents to recover the most fraud loss.
Cristine Marsh
Bio: Cristine is a machine learning scientist on the fraud team at Affirm. Prior to Affirm, she worked at IBM and studied at Galvanize.
Talk: “Who measures the metrics.” Cristine will go over metrics for imbalanced classes and how we determine which ones are most reliable for hyperparameter selection.
Fang Fang
Bio: Fang is a machine learning scientist at Affirm working on fraud prevention. She worked at Disney and studied at Yale before joining Affirm.
Talk: “Applying machine learning to new markets.” Fang will talk about how we use machine learning to expand our business to new markets where there is no previous data available.