Senior Computer Vision Engineer

About Trace: Trace is an elite applied AI infrastructure pioneer building the premier data marketplace for physical AI. While frontier language models have scaled exponentially on internet text collections, physical AI and robotics models remain heavily constrained by a lack of high-fidelity, real-world training datasets. Trace closes this loop by creating the hardware and software systems necessary to capture, align, and transform multi-modal sensors recording human physical work into enterprise-grade training data for embodied systems, robotics platforms, and automated machines operating outside the digital browser environment.

Position Overview

We are seeking a highly technical, high-agency Senior Computer Vision Engineer to take absolute ownership of the spatial perception layer inside our core data pipeline. In this load-bearing engineering role, you will build the mathematical and algorithmic systems that convert raw, sensor-heavy environment streams into flawlessly aligned, highly reliable spatial representations. Your day-to-day operations will span camera intrinsics/extrinsics calibration, multi-sensor time synchronization, and tracking diagnostics. This track is built for a pragmatist comfortable shipping spatial perception frameworks natively across real hardware structures where drift, noisy captures, and degraded tracking occur.

Key Responsibilities

• Spatial Layer Pipeline Design: Own and scale the spatial perception models that turn raw hardware captures into aligned datasets that downstream annotation, scene labeling, and policy training modules depend upon.
• Multi-Sensor Calibration: Design and implement automated scripts to manage camera intrinsics, extrinsics, sensor fusion alignments, and microsecond-level time synchronizations across custom capture rigs.
• Trajectory Recovery Tracking: Build, evaluate, and tune robust **SLAM, VIO (Visual-Inertial Odometry), and 3D mapping** pipelines to recover highly precise 6-DoF spatial tracking paths from physical captures.
• Field Failure Diagnosis: Debug and mitigate severe real-world production hardware faults, including localization drift, sensor misalignments, degraded tracking loops, and noisy multi-modal reconstructions.
• Ground-Truth Benchmarking: Formulate rigid benchmarking standards and metrics pipelines to quantitatively measure whether the spatial tracking layer is continuously scaling in accuracy.
• Applied Research Translation: Partner closely with Trace Labs (our internal research arm) to feed reliable 3D structural data into advanced semantic segmentation and custom pose estimation workflows.

Required Skills & Qualifications

• Strong professional background running computer vision engineering pipelines, with a deep focus on at least one vector: **SLAM, VIO, visual odometry, mapping, or localization**.
• Hands-on production tracking deploying camera calibration models, sensor fusion configurations, or real-time state estimation layers.
• A verified history of shipping functional perception architectures onto real hardware targets operating in wild environments (such as robotics, autonomous vehicles, AR/VR headsets, or drones).
• Comfortable debugging systems across software, sensors, and data quality parameters natively rather than evaluating abstract models in isolation.
• Outstanding technical communication and problem-solving skills, with a collaborative and emotionally mature working philosophy.
• Location Context: 100% remote-first operational infrastructure flexibility open to qualified engineers permanently based within the United States (with a strong preference for New York).

Preferred Strategic Indicators (Nice to Have)

• Practical familiarity with advanced spatial computing primitives, including 3D reconstruction, Structure from Motion (SfM), pose graph optimization, or bundle adjustment.
• Prior experience managing large-scale data ingestion or routing architectures tailored for multi-camera configurations, Spatial Computing arrays, or **LiDAR** arrays.
• An advanced academic background (Master’s or PhD) focusing on Computer Vision, Robotics Engineering, or related spatial fields.

What We Offer

• Targeted Annual Base Salary: $150,000 – $300,000 USD flat rate.
• Lucrative early-hire corporate equity options to directly share in the market cap upside of our physical AI marketplace.
• Profound remote-work operational independence supported by a tight-knit, expert engineering ecosystem composed of ex-founders and PhDs.
• The generational opportunity to design the core infrastructure determining how the next frontier of robotics and embodied AI models learn to perceive the physical world.