Founding Engineer

Founding Robotics Software and AI Engineer
Dream Robotics is building general-purpose robots that can perform useful work in businesses and homes. We are customer-centered, but we are also building core technology: multimodal sensing, scalable wearable data collection systems, robot learning infrastructure, and cost-efective mobile robots designed for real deployment.
This role is for someone who is strongest in software and AI, but is comfortable working on real robots. You should be able to train models, debug pipelines, read logs, SSH into a robot, reason through a control failure, and physically help get the system running again when a policy drives the robot arm into the table.
You will build the software systems that make our robots useful in the real world. That includes training infrastructure, offline data processing pipelines, model design, model evaluation, feedback control, computer vision, firmware, robot runtime systems, and deployment infrastructure.
What you’ll work onModel training for robot policies, perception models, and vision-language-action systemsOffline data processing pipelines for robot demonstrations, multimodal sensor data, video, tactile data, and robot logsModel evaluation tools, benchmarks, failure analysis workflows, and regression testingRobot control software, feedback control, motion execution, and safety-critical runtime systemsComputer vision, perception, SLAM, calibration, camera drivers, and sensor integrationNixOS-based robot software, networking, device management, and field debuggingFleet management, robot observability, remote diagnostics, OTA updates, and deployment toolingFirmware and low-level interfaces across cameras, actuators, sensors, embedded controllers, and compute systemsSoftware infrastructure that supports real customer deployments, not just lab demosYou do not need to have done every item above before. We care most about strong software ability, robotics judgment, fast learning, and comfort owning messy real-world systems end to end.
What we’re looking forMaster’s degree or PhD in robotics, computer science, electrical engineering, or AIStrong experience training VLA modelsStrong software engineering skills, especially in systems that interact with real hardwareExperience building software for robotics, autonomous systems, embodied AI, drones, AVs, manipulation, mobile robots, or similar physical systemsStrong Python and C++/Rust skillsComfort working in Linux environmentsExperience with robot data, model training, perception, controls, simulation, or deployment infrastructureAbility to debug across the stack: model behavior, logs, networking, sensors, drivers, control loops, and hardware interfacesGood engineering judgment around reliability, safety, latency, observability, and field deploymentWillingness to work hands-on with robots, including setup, testing, debugging, and occasional mechanical fixes
Helpful experienceVision-language-action models, imitation learning, reinforcement learning, diffusion policies, or robot foundation modelsLarge-scale model training, distributed training, data pipelines, dataset curation, or experiment trackingSLAM, visual-inertial odometry, calibration, mapping, localization, or navigationCamera drivers, sensor synchronization, embedded Linux, ROS 2, real-time systems, or firmwareRobot fleet management, OTA updates, remote monitoring, networking, or production deployment systemsSimulation, sim-to-real, synthetic data, or evaluation environmentsManipulation, mobile manipulation, humanoids, dexterous hands, tactile sensing, or multimodal sensing
Who you areYou are a software-first robotics engineer who wants to build systems that work outside the lab. You like models, but you also care about data quality, evaluation, deployment, reliability, and the unglamorous infrastructure that makes robots actually useful.You move quickly, communicate clearly, and take ownership. You are comfortable with ambiguity, but you do not hide behind it. When a robot fails, you want to know why, fix the system, and make sure the same failure is caught automatically next time.
Why joinYou will work on the full stack of embodied AI: data collection, model training, perception, control, deployment, and real customer use cases. You will help build robots designed to scale beyond demos into useful labor across businesses and homes.
CompensationBase salary $150k-200k. Equity 1-2%