Build the AI the earth runs on.
We hire researchers, engineers and geoscientists who want their work in production, deciding where a well goes, how a basin stores carbon, how a grid balances, not filed in a paper drawer.
A small team, distributed across time zones, shipping like a much larger one.
Research that ships
We publish, but the bar is production. If it can't survive a real well or a real grid, it isn't done.
Own the whole stack
Foundation model, agentic OS, applications. You'll see your work end to end, not one slice of someone else's pipeline.
Async, not always-on
Distributed across time zones with the handoff discipline to match. Deep work is protected, not interrupted.
Evidence over hype
We name our uncertainty and show our working. Rigour is the culture, not a compliance step.
Six disciplines under one roof.
Foundation models, segmentation, set-prediction and agentic reasoning for subsurface and energy data.
Multi-GPU on-prem training, orchestration, and the deployment fabric that keeps operator data inside the estate.
Interpretation, validation and the domain judgement that decides whether a model's output is geologically real.
From research bench to a tool an engineer trusts at 3am, and the field rollout that gets it adopted.
Bring the same discipline that de-risks a well to CCUS, solar, wind, battery and EV assets.
The interfaces and the Synapsa learning experience that put the platform in people's hands.
The best people here don't ask what the model gets. They ask what the drilling engineer with ten million dollars on the line gets.
Built for people who take the long view.
Remote-first, truly
Work where you do your best thinking. We've been distributed since before it was fashionable.
Real compute
On-prem multi-GPU and DGX-class hardware, your experiments aren't rationed.
Research time
Protected time to publish and to go deep, because the best product ideas start there.
Visa & relocation
For the right person we sponsor and help you move. Talent isn't bound by geography.
Intern-to-hire
A proven path from annotation and applied research into permanent research and engineering roles.
Equity
Meaningful ownership in the group building the AI the energy industry runs on.
Four steps. No trick questions. About three weeks.
Intro call
A 30-minute conversation about your work and what you want to build next. No CV theatre.
Deep dive
A technical discussion grounded in a real problem from our domain, not a whiteboard puzzle.
Work session
A paid, scoped exercise or portfolio review. We look at how you think, not how you memorise.
Meet the team
Meet the people you'd work with across time zones, ask anything, and we make a decision fast.
10 roles open now
Don't see your exact role? The best hires rarely match a posting. Tell us what you'd build and why it belongs at EarthScan.
