← Autonomy Labs
Autonomy Labs · Rover Program
Build With Us · open call
Eight kinds of work
Half need zero code
Peoria, IL · remote-friendly
DocCONTRIB-01 StatusPublic · v0.1 AuthorR. Kumar Updated2026-06-11
You don't need to be a coder.
You don't need to be a tech person.
You need will, and an intuition for how systems fit together.
We're building an autonomous rover that learns its world the way you'd teach a sharp new hire: walk it around, point at things, explain why. The architecture is published, the design logs are dated, and the whole thing is being assembled in public — by whoever shows up with creativity and persistence. Most of the work below has never been done before by anyone, which means experience matters less than the willingness to figure it out.

The work · eight kinds of minds

Firmware · hard real-time
The Safety Floor

The work: the microcontroller layer that must never fail — bump and cliff detection, emergency stop, the watchdog that takes over if the brain hangs. Deliberately dumb, geometric, and fast.

The mind: methodical. You read datasheets for fun and think "what's the failure mode?" before "what's the feature?"

Electromechanical
Making It Move

The work: motors, drivetrain, power, encoders, weatherproofing. Torque math for sloped Midwestern yards. The part of robotics that gets grease on your hands.

The mind: the garage builder. You've fixed things that weren't supposed to be user-serviceable.

AI models · edge compute
The Thinking Layer

The work: running vision-language models on a credit-card-sized computer and finding out what they can actually do. Benchmarks, prompts, quantization, honest measurement.

The mind: the experimenter. You'd rather run the test than argue about the answer.

Simulation · validation
The Proving Ground

The work: simulation environments, test design, the discipline of "how do we know it works before it's on someone's lawn?" NVIDIA Isaac Sim, scripted scenarios, real-image holdout sets.

The mind: the skeptic. Your favorite question is "yeah, but how do you know?"

No code required Human–machine interface
Teaching the Rover

The work: the training walk — how a person teaches the rover boundaries with gestures, voice, and a tablet. Designing the vocabulary of pointing at the world. Full protocol here.

The mind: the teacher. You're good at explaining things to people who don't share your context.

No code required Storytelling · documentation
Telling the Story

The work: design logs, demo videos, photography, the public record of a machine being born. We believe the dated paper trail of "said it in May, demonstrated it in August" is worth more than any pitch deck.

The mind: the storyteller. You notice the moment worth capturing while everyone else is staring at the terminal.

No code required Field operations
Yards & Logistics

The work: test-yard hosting (got a lawn with character — slopes, trees, a ravine? we want it), parts sourcing, 3D-print runs, field-day coordination. The rover needs real terrain and someone who makes things happen.

The mind: the get-it-done person. Your superpower is that things you say will happen, happen.

No code required Operations · funding
The Paperwork That Funds It

The work: grant writing, program ops, the unglamorous machinery that turns a garage project into a funded one. We're actively working a federal SBIR track — this work is load-bearing.

The mind: the operator. You find satisfaction in the document being airtight.

How it works · taste before you commit

01
Tell us what pulls you One or two areas from above — or something we didn't think of. Gut feel counts. You don't need to justify it with a résumé.
02
Get a small, real first task Weekend-sized, concretely scoped, with a deliverable. Not busywork — every taster task is something the project genuinely needs.
03
Keep what fits, drop what doesn't Try a second area if the first one didn't light up. The test isn't whether you finished — it's whether you kept thinking about it after you put it down.
Two honest notes. This is early-stage, unpaid, build-in-public work — the compensation is real ownership of a piece of something new, frontier tools in your hands, and a dated public record of what you built (if you want one). And credit is opt-in: some contributors want their name on the work, some prefer to stay quiet. Both are first-class here, no questions asked.
If any of this pulled at you, say so.
One or two sentences is plenty: which area, and what made you curious. We read everything and reply to everyone.
Prefer DMs? Find Rajeev on LinkedIn →