The Role
As Flight Controls / Embedded Robotics Lead, you will own the flight stack end-to-end—from estimator to controller to failsafes to the embedded plumbing that makes it robust in the real world. Your north star is operator trust: the drone should hold position, fly smoothly, and recover gracefully even when the environment is messy (low light, textureless surfaces, dust, GPS-denied spaces, tight corridors, multipath, airflow disturbances).
This is a high-impact role working directly with the founder and mechanical engineering cofounder. You’ll set the technical direction for stability, indoor navigation primitives (optical flow / depth / VIO as appropriate), tuning workflows, and “black box” logging so we can iterate fast and debug faster.
What You’ll Own
- Flight control performance: crisp, stable, low-overshoot control that feels locked-in indoors.
- State estimation / sensor fusion: IMU fusion with vision/depth aids as needed; drift reduction in GPS-denied environments.
- Indoor stabilization: optical flow, depth, VIO primitives, and the practical reality of lighting, texture, and dust.
- Failsafes & recovery: safe behaviors under loss of tracking, low battery, link degradation, sensor faults, or collisions.
- Tuning system: a repeatable, operator-friendly tuning process (not a wizard ritual).
- Reliability instrumentation: black-box logging, event markers, and performance telemetry to diagnose issues quickly.
- Embedded execution: real-time constraints, hardware integration with compute/FCU, and “it boots every time” robustness.
Early Outputs (First 60–90 Days)
You will ship tangible improvements fast. Example targets:
- Stable indoor hover mode with materially reduced drift and “hands-off” confidence.
- Low-drift profile for hallway/room operations, including smooth translation and stopping behavior.
- Repeatable tuning process (documented, testable, and not dependent on one person’s brain).
- Black-box logging (flight + estimator + control loops + fault events) with a simple retrieval + review workflow.
- A “monkey-flyable” control profile: predictable handling for stressed operators, not hobbyists.
Responsibilities
- Design and implement flight control algorithms (attitude, rate, position, velocity control as appropriate).
- Build and improve sensor fusion pipelines (EKF/UKF/complimentary approaches), integrating IMU + baro + magnetometer + vision/depth where useful.
- Evaluate and integrate indoor aiding sources (optical flow, depth cameras, VIO) with clear performance envelopes and failure modes.
- Develop failsafe logic and fault detection: sensor sanity checks, watchdogs, degraded-mode behavior, and safe land/return behaviors suitable for indoor ops.
- Create a rapid iteration loop: bench tests → tether tests → indoor test course → field-like scenarios.
- Build logging/telemetry pipelines that support debugging, regression testing, and performance baselining.
- Work closely with mechanical + electrical to ensure sensors are mounted, isolated, and placed for estimator health and RF sanity.
- Define acceptance criteria and test plans for “operator-ready” flight behavior.
What “Good” Looks Like
- The drone holds position indoors without constant stick correction.
- Drift and twitchiness are controlled and predictable.
- The system degrades gracefully (not catastrophically) when sensing gets weird.
- Tuning is documented, repeatable, and testable across builds.
- Every weird behavior can be diagnosed with logs—not vibes and guessing.
Required Qualifications
- 5+ years (or equivalent) building flight controls, robotics controls, or embedded autonomy systems.
- Strong fundamentals in control theory (PID, cascaded loops, LQR/modern control familiarity helpful) and estimation (Kalman filtering concepts in practice).
- Hands-on experience shipping embedded systems with real-time constraints (C/C++ and/or Rust strongly preferred; Python for tooling a plus).
- Proven experience integrating IMU-based estimation with additional sensors (vision, depth, optical flow, GNSS-denied localization, etc.).
- Practical debugging skill: oscillations, drift, estimator divergence, sensor noise, vibration coupling, timing issues.
- Comfort working in scrappy early-stage environments: you can prioritize, ship, and iterate.
Preferred Qualifications (Nice to Have)
- PX4 or ArduPilot experience (custom modules, estimator/controller modifications, tuning workflows).
- VIO/SLAM familiarity (you don’t need to be a PhD, but you understand failure modes and what “good enough” looks like).
- Experience with small multirotors, micro drones, or tight SWaP constraints (size, weight, power).
- Experience building “flight data recorder” / black-box systems and analysis tools.
- Background in defense / public safety / ruggedized robotics environments.
Tooling & Stack (Flexible)
We’re not religious about tools—we’re religious about results. Expect a mix of:
- Embedded C/C++ (flight stack + drivers), Python for analysis/testing
- Real-time logging + post-flight analysis pipelines
- Hardware-in-the-loop (HIL) and simulation where it speeds iteration
- Rapid prototyping + test rigs + indoor test course scenarios
Security / Compliance Note
This role may involve work related to public safety and defense-adjacent use cases. U.S. work authorization is required, and additional compliance requirements may apply as the program matures.
Why This Role Is Worth Your Time
- You own the technical heart of the product. If you nail this, we win.
- Real-world stakes: you’re building for people who need reliability, not gadget novelty.
- Small team, high trust, fast iteration, and direct founder access.
Location
Orange County, CA preferred, but we can support hybrid/remote for the right person with periodic on-site test weeks.
Compensation
Competitive salary + meaningful equity, commensurate with experience and fit. (We care more about shipping than title.)
