The Mission
GRAM is a self replication company creating populations of insectoids for the physical economy.
Our first research frontier is self-preservation: the base case of physical self-replication. Our machines will survive, coordinate, and recover without humans. We believe scalable machine labor requires more than single-agent task generality or machines shaped in our image.
About the role
You will own the systems and methods that generate high-quality robot data from physical work, spanning demonstration, teleoperation, intervention, autonomous operation, instrumentation, operator tooling, scenario execution, and source-quality control. You will translate capability gaps into collection campaigns and use measured downstream results to decide what the machines should experience next.
This is a senior engineering role responsible for how experience is produced and validated at capture, including meaningful variation, failures, recoveries, and episode-level evidence. You also define the capture contracts that keep recorded experience usable for reproducible datasets and replay. Success means each campaign yields data that can change a training or evaluation decision—not operating hours without learning value.
What you will do
- Build field-capable demonstration interfaces and robust systems for teleoperation, intervention, autonomous rollout, and recovery-data collection in realistic physical work.
- Translate model failures and evaluation gaps into controlled scenarios, collection protocols, sampling priorities, and measurable acceptance criteria.
- Design operator interfaces with explicit command authority, latency budgets, feedback, safe handoff, and emergency behavior.
- Instrument demonstration tools, robots, operators, and environments so perception, action, timing, contact, intervention, and task outcome remain aligned and valid at capture.
- Commission collection stations and diagnose failures spanning sensors, controls, networking, operator input, robot execution, and recorded data.
- Define operator procedures, calibration checks, training, and escalation paths that produce consistent evidence across people and sessions.
- Measure whether collected experience changes model or system performance, then use the result to refine the next campaign.
Minimum qualifications
- Bachelor's degree in computer science, electrical engineering, mechanical engineering, robotics, or a related field, or equivalent practical experience.
- Strong C++ or Python programming ability in Linux, including experience with robot middleware, sensor streams, command interfaces, and hardware debugging.
- Direct experience building and repeatedly operating a teleoperation, shared-control, portable demonstration, instrumented task-tool, or robot-data collection system for physical work.
- Demonstrated ability to characterize end-to-end latency, timing, calibration, command safety, and data quality using instrumented tests.
- Experience converting a learning, test, or capability objective into a physical collection protocol whose data produced a measured change in model or system performance.
Preferred experience
- Portable demonstration interfaces, instrumented task tools, VR, motion capture, haptics, retargeting, remote robot operation, or shared autonomy.
- Imitation learning, reinforcement learning, active learning, or failure-directed data collection.
- Networked real-time systems, video transport, time synchronization, or edge data capture.
The annual base salary range for this San Francisco Bay Area position is $170,000–$220,000. Health coverage, benefits, and generous equity come with the role. This role is on-site in the San Francisco Bay Area. A relocation bonus is available. We hire to start as soon as possible. We review what you’ve shipped; if it’s the caliber we seek, interviews take about a week, start to finish. We treat your work and conversations with discretion.
