Senior Machine Learning Engineer, Perception and Controls

Capstan Medical•West Menlo Park, CA

24d•Hybrid

About The Position

Capstan Medical is hiring a Senior Machine Learning Engineer for its robotics and controls team to develop perception for a robotic platform used in minimally invasive heart valve treatment. The role focuses on helping the robot determine the location of its devices inside the patient using intraprocedural imaging, including fluoroscopy and echocardiography. The engineer will develop image-to-robot registration and pose-estimation methods and bring them from research into a clinical system. This is a hands-on role building machine-learning components for a real-time robotic system. The work informs what physicians see during procedures and must meet real-time latency requirements and FDA design controls. The engineer will also shape the collection, organization, and use of robot and imaging data to improve device models and controllers, and will help establish the team's ML tools, standards, and priorities. The role is hybrid, with onsite work 3–4 days per week split flexibly between Santa Cruz and West Menlo Park, California, including at least one day per week in Santa Cruz. The position reports to the Robotics & Controls Manager.

Requirements

Master's degree in Computer Science, Electrical Engineering, Robotics, or a related field.
At least 4 years of professional experience in computer vision or machine learning, classical or learned, including at least one system shipped to production on real-world data.
Experience working with scarce, noisy, or expensive data, using synthetic data, augmentation, self-supervision, or physics priors to address data limitations.
Experience building ML using data from physical systems such as robots, cameras, medical imaging, or sensors.
Strong background in classical and learned computer vision, especially pose estimation, segmentation, tracking, or 2D/3D registration.
Solid understanding of rigid transforms, rotations, camera projection models, and calibration.
Strong Python and PyTorch skills and comfort reading and writing C++.
Ability to build reusable data pipelines, datasets, and evaluation harnesses.
Ability to assess research ideas and turn promising approaches into working prototypes.
Comfort with ambiguity and ability to break unclear goals into experiments, choose a direction, and own problems end to end.
Excellent communication skills, including explaining ML to controls engineers, clinical engineers, and physicians, documenting work, and helping the team determine when a model is trustworthy enough.
Ability to work onsite 3–4 days per week, flexibly split between the Santa Cruz and West Menlo Park offices, with at least one day per week in Santa Cruz.
Ability to thrive in a fast-paced environment and take ownership of problems end to end.

Nice To Haves

PhD in a related field.
Ultrasound or echocardiography experience; described as strongly preferred.
Fluoroscopy or X-ray experience, including 2D/3D registration or digitally reconstructed radiographs.
Experience tracking catheters, needles, surgical tools, or continuum robots in images.
Experience with sensor fusion and state estimation, such as Kalman filters or factor graphs.
Experience with system identification or learned dynamics for control, including residual models on top of physics.
Robotics software experience with ROS/ROS2 and on-device inference under latency constraints, such as ONNX or TensorRT.
Data platform and MLOps experience, including dataset versioning and experiment tracking.
Medical device or safety-critical experience, including validating ML under FDA design controls.
Experience as an early ML engineer on a team, establishing tools, data practices, and evaluation processes.
Publications at venues such as MICCAI, IPCAI, CVPR, ICRA, or IROS.

Responsibilities

Develop methods to register intraprocedural imaging, such as fluoroscopy and echocardiography, to the robot.
Estimate the pose of catheters and implants from medical images.
Combine image-based estimates with robot kinematics and other sensing, and quantify uncertainty in each estimate.
Build data pipelines to capture, align, label, and version imaging and robot data for training and evaluation.
Use simulation and synthetic data when real clinical data is limited.
Fit and improve models of device behavior from data and work with controls engineers to use them to improve control.
Work with clinical engineers and physicians to understand image content, procedural priorities, and acceptable accuracy.
Review research literature, prototype promising approaches, and evaluate their performance on the company's data.
Deploy models on the robot and meet real-time latency budgets.
Write validation materials, test datasets, and design documentation to support learned components under FDA design controls.
Help shape the team's ML tools, standards, data practices, and problem priorities, and provide ML guidance to the team.