Cofounder, Head of Field Autonomy

COFOUNDER, HEAD OF FIELD AUTONOMY — PLUTON Part-Time To Start · Remote + Field Deployment · Pre-Seed Stage · Equity As Compensation US WORK AUTHORIZATION REQUIRED — CITIZEN OR GREEN CARD ONLY, NO VISA SPONSORSHIP ABOUT PLUTON: Pluton Resource Intelligence is building a closed-loop probabilistic command system for mineral exploration. Four components, one architecture:

• NOUS

— the probabilistic belief engine. It fuses geological evidence into prospectivity scores with explicit uncertainty and calculates which survey or drill hole would tell us the most. Validated across Nevada: more than 80% of known gold fell in the top 5% of ranked ground within target districts, and on unlabeled public data it surfaced the Walker Lane belt and the Carlin Trend. The model updates in hours.

• CARTA

— the 3D command interface. It shows ranked targets with confidence levels, drill-hole recommendations with full specifications, and drone-survey blocks ranked by value of information.

• GAIA

— the simulation and mission-planning environment, where missions are rehearsed against synthetic terrain before they fly.

• ROBOTS

— EAGLE, our drone, flies today. BEAR, our ground platform, comes next. We are a Delaware C-corp partnered with operators and governments across the Middle East and Africa, with an active project in Saudi Arabia and signed partnerships with Saudi field-execution companies. We are actively raising our pre-seed round, with close targeted for Q1 2027. The team : two founders (CEO and CTO), a Head of Geoscience who built the model (previously the first data engineering hire at a wildfire-detection AI company, a PhD reaearcher from NASA with data systems work for USGS and NOAA projects), a PhD Robotics Engineer and patented inventor, and a Chief Exploration Geologist with 15 years running exploration campaigns in West Africa. Advisors include a former Rio Tinto copper exploration principal and a two-time AI founder and former VC. MISSION: Critical minerals are civilizational infrastructure. The world needs far more copper, lithium, nickel, rare earths and uranium than the exploration industry knows how to find. The deposits are there; the industry has not learned how to structure belief and allocate capital under deep uncertainty. Exploration is not a data problem. It is a decision problem. A single deep hole costs $500K–$2M, and less than 1% of greenfield targets become producing mines. Before capital goes into the ground, the questions that matter are which measurement will most change the decision and how fast it can be collected. At Pluton, the model picks the measurement and our robots collect it. THE ROLE: You own Pluton's field autonomy: the system that turns the model's recommendations into missions our robots fly, and returns clean data to the model. EAGLE flies today. Your job is to take it from a flying drone to a closed loop: NOUS picks the measurement, EAGLE collects it, NOUS updates, and the next mission follows. As BEAR comes online, it joins the loop you build. This is a cofounder seat. You are the architect and builder of the loop, and the person in the field when it first runs on a client's licence area. WHAT THE LOOP DOES: NOUS ranks candidate surveys by how much each would reduce uncertainty about the ore, and CARTA shows the top-ranked survey blocks. Your mission software turns a survey block into a flyable mission (line spacing, direction, terrain-following altitude, battery-split flights, geofences) and rehearses it in GAIA. EAGLE flies it with a geophysical sensor, starting with a magnetometer, on RTK-grade positioning. A pilot in command approves each flight. Data is quality-checked before the crew leaves the site, then processed automatically with a noise estimate attached. NOUS updates, and the next mission is generated. THE TWO HARD TECHNICAL PROBLEMS: Closing the loop in hours. In conventional exploration, getting from a survey recommendation to processed data and a new decision takes weeks or months. The target here is hours, so every manual step between the model and the aircraft has to be engineered out. Clean geophysical data from a drone. Motor and speed-controller noise, the airframe's own magnetic signature, heading effects, and timing offsets between the sensor and GNSS all show up in the data unless the integration accounts for them. A record with unmodeled platform noise is worse than no record, because the model takes it as evidence. **WHAT WE'RE LOOKING FOR: Must have:** US citizen or green card holder (we cannot sponsor visas) 5+ years building and deploying drone or field robotic systems, with time in the field under real conditions, not only lab demos Mission automation: you've written code that generates missions and uploads them to an autopilot, rather than planning each flight by hand in a ground-station app Sensor integration on drones: you've mounted, powered and logged a scientific sensor, dealt with motor and electrical noise, synced its clock to RTK/PPK GNSS, and run calibration flights Flight operations: you hold an FAA Part 107 certificate or will before the first flight, you've configured and tuned an autopilot, and you can repair motors, speed controllers, wiring and batteries in the field Automated data pipelines in Python: field checks that run right after a flight, and scripted processing from raw logs to analysis-ready data Agency: you diagnose what's wrong, decide what to build, and execute without being managed step by step Highly valuable: Geophysics: survey design, or magnetic data processing (corrections, leveling between lines, gridding) Planning under uncertainty: active sensing, Bayesian experimental design, or POMDPs ROS 2, for ground robots joining the same mission and QC software Cloud pipelines and APIs (AWS or similar) Beyond-visual-line-of-sight approvals or Part 107 waivers Export-control (EAR/ITAR) review experience; review of our platforms is part of the work Comfort in remote, infrastructure-limited environments (45°C heat, dust, generator power, limited comms) The stack: Python, Linux, Git, MAVLink/MAVSDK with PX4 or ArduPilot (or the DJI SDK), RTK/PPK GNSS workflows, Isaac Sim for mission rehearsal in GAIA, Oasis montaj or Harmonica, Verde and SimPEG for geophysical processing, AWS. WHAT THIS IS NOT: A research role. The output is deployed in the field, not published. A drone-pilot-for-hire role. You build the system; flying it is one part of the job. A vehicle design role. EAGLE exists; the work is the loop around it. A team with established process. You set the foundation others will build on. WHAT WE OFFER: Cofounder title and meaningful equity from day one Remote-first with field deployments (all travel and on-site expenses covered) Direct access to the founders, with no layers Full-time salary as we close our pre-seed A drone that flies today and a model that already ranks where it should fly next Ownership of the loop that runs Pluton's system in the field HOW TO APPLY: Apply through LinkedIn with your resume plus: A brief intro: who you are and why this role interests you to [email protected] Links to work you're proud of: GitHub, papers, field deployments One drone or field robotic system you built that collected real data, and what went wrong with the data the first time you took it outside If this listing has closed by the time you find it, apply through our careers page: https://www.pluton.earth/careers We respond to every application. No recruiters, please.