Machine Learning Engineer

The self-employed workforce is a rapidly growing, resilient, and colorful 60 million Americans. But self-employment comes with its own set of challenges: navigating taxes, accounting, bookkeeping, and business banking are just a few. That’s where we come in.

Found is building tools that give self-employed people the security and peace of mind that has historically only been possible at big corporations. We’re a business bank account that automates taxes and expense tracking because we believe small business owners should spend more time doing what they love and less time on their business finances.

We’re looking for kind, resourceful, and passionate people to join us in building the safety net for self-employment.

About this role: 

Thanks for your interest in Found! Our Data team is a full stack data organization of highly skilled analysts, data engineers, and MLEs. We’re responsible for everything from researching, training, and deploying ML models to detect and stop fraud in real time, to using data to inform the direction that our product teams should take / what to build next.

We’re a small but experienced team that has worked on diverse and complex problems at companies like Uber, Netflix, and Spotify. We’re excited to bring on another member to our team to enhance our Data Science and ML capabilities. You’ll start by partnering with our risk team to develop models to better identify risky activity, underwrite users for things like higher account limits or access to check deposits, and improve our understanding of user quality through identifying signals of ‘real’ vs. fraudulent businesses. We’re looking for someone who’s curious, proactive, and who is excited to dive in and immediately play a critical role impacting company-level strategic priorities. 

Some recent team accomplishments include: 

• Researching, training, and deploying an XGBoost model that predicts a user’s LTV to better shape our ad targeting and user acquisition strategy. 

• Deploying a new feature store that improved time to train and deploy models by ~50%. 

• Optimizing our batch and streaming jobs to improve reporting latencies from a matter of days to a matter of hours. 

• Training and building several models that monitor financial transactions in real time to make decisions on whether or not they should be allowed through or not.  

• Customizing onboarding flows based on user behavior and expected value to cut onboarding CAC by ⅓ and optimize upsells/CTAs.

Day to day, you will:

• Design, develop, and deploy machine …