Senior / Staff Product Engineer, Machine Learning

Radar•New York, NY

1d•Onsite

About The Position

We're looking for Product Engineers to build machine learning-based systems into core Radar products. This is a product-oriented role building new machine learning based systems into our backend, data infra, and mobile SDKs. The ideal engineer for this role is someone who is primarily an ML engineer but wants to broaden their skills into other stacks like server, data and mobile. The perfect candidate will see themselves as generalist who has built real ML systems and is ultimately motivated by driving impact to products and customers by building end-to-end features that leverage machine learning. We have many ML challenges across our Geofencing, Maps and Fraud products. How we work: Most of our engineering team are former technical co-founders or former Radar interns from schools like Waterloo and CMU. Most engineers at Radar fit one of two molds, technically: either Staff level expertise in one stack, or "Multi-Stack" at any level. We say "Multi-Stack" because "Full-Stack" has the connotation of "Frontend and Backend", but Radar Engineers might also work on Mobile or Data engineering. Not that you need to be an expert in all of those, but a desire to learn, jump around to different stacks, and get things done is the important part. We care a lot about shipping fast and talking to customers. We're committed to our product vision of full-stack location infrastructure, but we also know that customer feedback is a treasure map to gold. Even though Slack is the brain of our company, working together in-person in our NYC HQ is the fastest way for us to get things done. We meet on Mondays to plan out work for the week in small groups and use Linear for planning. To us, a week is a long time, and we expect to ship big things every week. The stack: We have systems that leverage lightgbm and random forests using scikit and Rust and we need to build out new systems impacting additional products. The server is a TypeScript Node.js app and a Geospatial Rust database we built call HorizonDB. We use MongoDB, S3/Athena, Redis, Airflow and everything is deployed to AWS. Most engineers are in the on-call rotation. We sponsor OpenStreetMaps, MapLibre, and OpenAddresses. How we use AI: Engineers choose what AI tools they use, Claude being the most popular. We're actively building Claude skills - for example we've taught it how to debug HorizonDB, our geospatial database. All code changes are reviewed by an Engineer knowledgeable in that area. Claude and Copilot also review all PRs. There is a range of how much engineers use AI. Most use it daily if not weekly. We are excited about what AI can do, but we also recognize the risks and don't compromise our coding standards. The hiring process: After a call with our Technical Recruiter, you'll do several technical Zoom calls with members of our engineering team: code screen, coding round, and system design round. If those go well we'll invite you to our NYC HQ for a final round interview. You'll meet one of our co-founders, someone from outside engineering, and meet more people from Radar. We'll go into more depth about how we work to see if there is a match.

Requirements

Have experience building machine learning based products in production at scale
Don't think of yourself as an "ML Engineer"
Are interested in talking to customers or prospects and making them successful
Are deeply curious about how things work, and have the tenacity to sit with hard problems and power through them

Nice To Haves

Are a former technical co-founder
Have experience with anomaly detection, indoor positioning, or mapping infrastructure

Responsibilities

Work on core Radar ML infrastructure built with Python, Rust, Airflow, Spark and new systems you build
Build new systems for our Fraud products: anomaly detection, user and device risk scores, device fingerprinting, and emerging threat vectors
Build new systems for our Maps products: search ranking and query classification for addresses, points of interest and others leveraging "learn-to-rank" systems, LightGBM, and new systems you build.
Create road traffic models based on historical and live data to improve the ETA accuracy of our routing engine.
Leverage AI to ingest address and point-of-interest data from across the web.
Build new systems for our Geofencing products: indoor positioning leveraging mobile device sensors (BLE, barometer, motion, ultra-wideband, etc)
Work on full-stack features across several of backend, data infra, mobile, and frontend
Push the limits of what is possible with location services on iOS and Android
Have your work run on 300M+ devices
Talk to Radar customers and prospects, hear their feedback, incorporate it into your work, and make them successful