Most ML roles ask you to fine-tune someone else’s idea. This one asks you to build the thing that changes how professionals work entirely.

We’re hiring on behalf of a VC-backed SaaS company building browser autopilot – AI that runs directly through professionals’ existing browsers, automating the repetitive work that consumes their day. No latency. No migration. No IT friction.

Backed by Anthropic.

The problem is hard, the team is exceptional, and the window to be a founding engineer on something like this doesn’t stay open long.

What you’ll work on

You’ll own the ML systems that make autonomous browser interaction actually work – not as a feature team, not from a ticket queue, but as a core technical decision-maker from day one.

Concretely, that means:

• Building and training the models that understand browser state, user intent, and workflow context in real time
• Designing the inference architecture that delivers instant responses through live browser sessions – latency here isn’t a nice-to-have, it’s the product
• Developing the systems that learn from user behaviour over time – making the autopilot genuinely smarter the more it’s used
• Working at the intersection of LLMs, browser automation, and enterprise productivity tooling – a genuinely novel problem space with very few established playbooks
• Owning decisions end-to-end: data pipelines, model architecture, eval frameworks, and production deployment

Your commits ship to real users. Your architecture calls shape the product roadmap. This is the kind of role where research and engineering are the same job.

Who we’re looking for

Top 10% graduates from top 20 CS programs. Phi Beta Kappa honorees or equivalent. The kind of person who treated their thesis like a startup and left their internship with a paper idea.

You’ve already shipped something in ML – at least one internship or a year in industry. You know what production looks like. You’ve felt the difference between a model that works in a notebook and one that holds up under real load.

You likely have:

• A CS degree from a top-tier program – MIT, Stanford, CMU, Berkeley, Caltech, Cornell, UIUC, or equivalent
• At least one ML internship or 12+ months in industry
• Deep hands-on experience with PyTorch and modern transformer architectures
• End-to-end comfort across data, training, eval, and deployment
• Genuine obsession with agents, browser automation, or the future of human-computer interaction

What you get:

• $200,000 – $300,000 base + founding team equity – own a real piece of what you build
• Full medical, dental, and vision for you and your family
• Breakfast, lunch, dinner, snacks, and drinks on-site daily
• A beautiful office in the heart of SF with a team that actually wants to be there
• Direct Anthropic backing – frontier model access and research relationships most companies will never have