Forward Deployed Robotics Engineer

Gecko Robotics•Pittsburgh, PA

26d•Hybrid

About The Position

Gecko Robotics is seeking a highly skilled Robotics Engineer specializing in underwater sensor fusion and localization to lead the development and deployment of robust navigation systems for our hull-cleaning and NDT inspection robots. The ideal candidate will have deep expertise in fusing data from underwater sensor suites - including IMUs, DVLs, pressure sensors, acoustic positioning systems such as USBL/LBL, and sonar-based relative sensing - to produce accurate and reliable state estimates in GPS-denied, close-to-structure environments. You will be responsible for solving real localization problems in the field: compensating for drift, handling intermittent or delayed aiding, managing sensor dropouts and acoustic outliers, and delivering pragmatic solutions that can be deployed under tight timelines. To be successful in this role, you must possess strong hands-on experience designing and implementing advanced filtering algorithms for real-world robotic systems, such as error-state Extended Kalman Filters (EKF), Unscented Kalman Filters (UKF), and Particle Filters where appropriate. You should be adept at deriving and maintaining both linear and nonlinear state-space models to continuously estimate critical variables such as position, velocity, attitude, and sensor biases in challenging underwater conditions. Additionally, the role requires expertise in calibrating and validating real sensor stacks - including IMU bias estimation, DVL alignment and dropout handling, pressure-depth offsets, acoustic latency, and timing synchronization - as well as a strong understanding of underwater navigation concepts such as dead reckoning, inertial navigation, observability in constrained motion, degraded-mode localization, and robust fusion when absolute aids are sparse or unreliable.

Requirements

Proficiency designing and implementing localization and navigation filters, including Kalman Filters, error-state Extended Kalman Filters (EKF), Unscented Kalman Filters (UKF), and Particle Filters.
Experience fusing data from underwater sensors, including IMUs, DVLs, pressure sensors, USBL/LBL systems, sonar-derived relative measurements, and other non-vision localization inputs.
Strong background in formulating dynamic models for underwater vehicles, including position, velocity, attitude, sensor biases, and other latent states.
Understanding of observability under constrained motion and close-proximity operations.
Expertise in managing real-world uncertainties such as drifting IMU biases, DVL misalignment or bottom-lock loss, pressure sensor offsets, acoustic multipath and latency, clock synchronization issues, and magnetic disturbances.
Thorough understanding of GPS-denied underwater localization concepts, including dead reckoning, inertial navigation, acoustic aiding, hull-relative localization, and fallback behavior.
Demonstrated ability to deploy working sensor fusion solutions within tight timelines, debug failures from field logs, tune estimators, and make practical engineering tradeoffs.
Familiarity designing resilient localization architectures that can tolerate delayed and asynchronous measurements, intermittent aiding, sensor dropouts, and compute or bandwidth constraints.

Responsibilities

Lead the development and deployment of robust navigation systems for hull-cleaning and NDT inspection robots.
Fuse data from underwater sensor suites (IMUs, DVLs, pressure sensors, acoustic positioning systems, sonar-based relative sensing) to produce accurate and reliable state estimates in GPS-denied, close-to-structure environments.
Solve real localization problems in the field, including compensating for drift, handling intermittent or delayed aiding, managing sensor dropouts and acoustic outliers.
Deliver pragmatic solutions that can be deployed under tight timelines.
Design and implement advanced filtering algorithms for real-world robotic systems (e.g., EKF, UKF, Particle Filters).
Derive and maintain linear and nonlinear state-space models to estimate position, velocity, attitude, and sensor biases.
Calibrate and validate real sensor stacks, including IMU bias estimation, DVL alignment and dropout handling, pressure-depth offsets, acoustic latency, and timing synchronization.
Apply understanding of underwater navigation concepts (dead reckoning, inertial navigation, observability, degraded-mode localization, robust fusion).