Computer Vision Engineer

Senior Geometric Computer Vision EngineerRemote (US) · Full-time · PST hours$155K–$205K base + strong Incentive Stock OptionsUS citizen or current green card holder requiredThe companyThe company builds VIDAS™, a passive-sensor perception stack that turns video from automotive-grade cameras and IMU data into dense, semantic 3D representations in real time.It is positioned as a fully redundant alternative to LiDAR and radar, so the bar is deployment reliability, not demo quality.The business is already generating revenue and shipping to customers across automotive, agriculture, healthcare, and defense. It is backed by Khosla Ventures, Two Sigma Ventures, the US Army, and other leading investors.The team is about 20 people in the US, and the company was founded in 2015. In that setting, the person in this seat will shape core technical decisions, not just implement them.The roleThis is a senior IC role for an engineer who wants to own the geometry and estimation layer of the product.You will work at the center of VIDAS: pose estimation, state estimation, camera-IMU fusion, calibration, 3D reconstruction, and the evaluation machinery that tells the team whether the system is actually improving.The work is hands-on. Expect to spend your time on math, code, experiments, and failure analysis, with enough scope to influence architecture and mentoring as the team grows.The technical problemPassive-sensor perception fails in specific ways: timing drift, calibration error, noisy IMUs, motion blur, occlusion, degenerate geometry, and deployment conditions that do not match lab data.The hard part is not finding a paper for each subproblem. The hard part is knowing which assumption broke, which part of the pipeline needs to change, and how to get the system back to a stable real-time state estimate.What you'll own• Pose estimation and SLAM: design and advance algorithms for motion estimation and consistent 3D structure from camera and IMU data.• Visual-Inertial Odometry: improve the core VIO stack so it remains stable, accurate, and usable in production systems.• Sensor fusion: own pipelines that combine camera and IMU signals in lidar-free deployments.• Classical estimation: build and tune EKF and other state-estimation approaches where they are the right tool.• Multi-view geometry and calibration: develop calibration systems and geometric tooling that keep the rest of the stack honest.• 3D reconstruction: improve dense and semantic reconstruction methods that downstream systems can trust.• Real-time perception: keep the system within latency and reliability bounds for autonomy and robotics use cases.• Failure analysis: inspect output quality, isolate error sources, and turn findings into specific algorithmic improvements.• Technical leadership: help raise the bar through design reviews, code reviews, mentoring, and clear technical judgment.Who this is forYou are likely a strong fit if you have:• 3–10 years of experience in computer vision, robotics perception, or a closely related field.• Deep familiarity with geometry, probability, linear algebra, and state estimation.• Built or debugged SLAM, VIO, calibration, or sensor-fusion systems that had to work outside a paper or demo.• The habit of working back to first principles when the literature is incomplete or the system fails in an unexpected way.• Strong mathematical judgment that shows up in code quality, evaluation design, and debugging speed.• Comfort analyzing failure modes, not just tuning average metrics.• Experience turning ambiguous requirements into production behavior without needing heavy process around you.• The ability to explain technical tradeoffs clearly to other engineers and non-specialists.• Willingness to work on defense-related projects.• Eligibility to work in the US without visa sponsorship.Why nowThe company already has customers and revenue. The next bottleneck is not proving that the approach works in principle; it is making the core perception stack more accurate, more stable, and easier to deploy across customer environments.That is the point where a senior geometry engineer has outsized leverage.This role is not for you if• You want to work only on model training or experimentation without owning production behavior.• You prefer problems where the answer is already cleanly solved in the literature.• You are uncomfortable working on defense-related projects.• You need visa sponsorship.• You want a large organization where this subsystem is split across many teams.Compensation and logistics• Base salary: $155K–$205K• Equity: strong Incentive Stock Options• Location: remote across the US• Hours: expected to work in PST• Travel: up to 5%• Office: office space available in Redwood City, CA• Visa sponsorship: not available• Eligibility: must be a US citizen or current green card holder• Employment: full-timeBenefits include comprehensive health care plans, 401(k) with matching, flexible schedules, and discretionary PTO.About AuroraAurora helps exceptional engineers find the right role at some of the most ambitious startups worldwide.We work with teams that value high ownership, technical rigor, and clear scope.