Senior ML Infrastructure Engineer

About Arlo

Arlo is rebuilding health insurance for small businesses from first principles. Most American healthcare costs stem not from medical care itself, but from the machinery surrounding it: middlemen taking cuts, undetected fraud, and billing systems designed to fight over payment rather than deliver care. Arlo eliminates these inefficiencies by identifying fraud earlier, steering members toward higher-quality and lower-cost care, automating operational overhead, and removing vendors whose primary function is extracting value.

AI is the foundation enabling this transformation. Arlo deploys machine learning across underwriting, operations, clinical programs, and member experience. The company operates at meaningful scale: profitable, managing hundreds of millions in premiums, covering tens of thousands of members, and growing rapidly through brokers, employers, and partners. Backed by Upfront Ventures, 8VC, and General Catalyst, with leadership from Palantir, YC companies, and veteran healthcare operators.

About the Role

As Senior ML Infrastructure Engineer, you'll build and own the infrastructure powering Arlo's underwriting engine—the core of the business. You'll architect systems trained on tens of millions of patient records and hundreds of millions of claims rows, then serve real-time quotes in seconds against inference-time datasets reaching trillions of rows. This role combines ML infrastructure ownership with hands-on opportunity to collaborate with data scientists and actuaries, testing and evaluating your own ideas alongside supporting theirs.

What You'll Work On

Training Infrastructure for Underwriting

• Build and own the infrastructure layer powering underwriting models trained on tens of millions of patients and hundreds of millions of claims rows
• Ensure training is reliable, reproducible, and scalable as data volume and model complexity grow

Real-Time Inference for Quoting

• Build and own the API layer producing quotes in seconds, serving trained models against trillions of rows of claims data across hundreds of millions of people
• Own latency, reliability, and scalability of the inference pipeline

Data Science Tooling

• Build tooling enabling data scientists and actuaries to iterate faster and more effectively
• Support experimentation and model evaluation workflows