Vision Systems Engineer

Exploration Technology Group; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Exploration Technology Group is seeking a hands-on Vision Systems Engineer to own the detection, tracking, and target discrimination software for a space-based IR sensing program. You will develop real-time pipelines on embedded hardware, integrating with simulations and hardware-in-the-loop tests.

The role emphasizes Kalman filters, multi-target tracking, CNN classifiers for discrimination, and MATLAB/Python/C++ development for production deployment in demanding space environments.

Qualifications

• Bachelor’s degree in a technical field; advanced degree preferred for senior roles.
• 5+ years in IR sensor systems or space-based imaging, defense experience preferred.
• Strong background in state estimation, tracking, and data association.

Responsibilities

• Own vision software stack for autonomous terminal guidance on embedded space hardware.
• Develop detection, tracking, state estimation, and discrimination pipelines.
• Train lightweight CNN classifiers for real-time embedded inference.
• Integrate algorithms into IR scene simulations and hardware-in-the-loop validation.
• Collaborate with electronics teams on ROIC interfaces and high-speed data paths.

Skills

Kalman filters

Multi-target tracking

Data association

Real-time embedded inference

Image processing

Python programming

C++ production deployment

Machine learning inference

Education

Bachelor’s degree in Software/Electrical/Optical/Physics/Aerospace Engineering

Advanced degree preferred

Tools

MATLAB

Python

C++

FPGA

GPU acceleration

Job description

About the Role

We’re hiring a hands‑on Vision Systems Engineer to own the detection, tracking, and target discrimination software stack for a space‑based IR sensing program supporting US national security missions. Your job is the vision algorithms: turning LWIR imagery into high-confidence track and guidance outputs in real time, on embedded space hardware, in demanding engagement environments.

What You’ll Do

You will own the vision software stack that enables autonomous terminal guidance for a national security space mission. This means building and validating the full pipeline from detection through tracking, state estimation, and discrimination, running on embedded space hardware under hard real‑time constraints.

• Target Detection and Clutter Suppression: Design and implement adaptive detection algorithms against Earth-background and cold-space clutter, including background estimation, non‑uniformity compensation, and track‑before‑detect approaches for high‑velocity terminal engagement geometries.
• Multi‑Target Tracking and State Estimation: Design and implement multi‑target tracking systems — track initiation, maintenance, correlation, and termination — using Kalman filters (EKF, UKF), IMM filters, and advanced data association methods (JPDA, MHT), generating the full terminal guidance output suite: LOS angles and rates (Az/El) for proportional navigation, target position and relative velocity, range and range rate, predicted intercept point, track confidence and lock flag, and discrimination confidence for GNC handoff.
• Target Discrimination: Develop and train lightweight CNN classifiers for target vs. decoy discrimination using radiometric imagery, optimized for real‑time embedded inference via quantization, pruning, or equivalent techniques, with full synthetic and real data pipelines for training and validation.
• Simulation Integration and Validation: Integrate algorithms into an existing end‑to‑end IR scene simulation framework, develop thermal signature models for synthetic training data, conduct Monte Carlo analysis and hardware‑in‑the‑loop validation, and own performance verification against program requirements.
• Signal Chain & Electronics Integration: Define detector readout architectures, work with analog and digital electronics teams on ROIC interfaces, low‑noise signal chains, and high‑speed data paths to on‑board processors
• Image Processing & Embedded Algorithms: Develop real‑time image processing pipelines — non‑uniformity correction, bad pixel replacement, background suppression, track‑before‑detect — on embedded FPGA and GPU platforms
• Performance Modeling & Test: Build end‑to‑end radiometric and noise models (NEDT, D*, SNR, clutter metrics), develop test plans, and execute lab and field characterization campaigns.

Qualifications

Education:

• Bachelor’s degree in Software Engineering, Electrical Engineering, Optical Engineering, Physics, Aerospace Engineering, or related technical field. Advanced degree preferred for senior applicants.

Experience:

• 5+ years of professional experience in IR sensor systems, electro‑optical payloads, or space‑based imaging, preferably in defense, missile defense, or national security space programs

Technical Skills:

• Strong theoretical and practical experience with state estimation: Kalman filters (EKF, UKF), multi‑target tracking concepts, and data association (JPDA, MHT, nearest neighbor)
• Proficiency in algorithm development in MATLAB and/or Python with demonstrated experience transitioning to C++ for production deployment
• Experience training and deploying CNNs or lightweight ML classifiers for real‑time embedded inference
• Experience with modeling, simulation, and testing workflows including Monte Carlo analysis, sensor data playback, or hardware‑in‑the‑loop environments
• Proficiency in radiometric modeling and performance analysis (NEDT, D*, MTF, SNR, background‑limited performance)
• Signal processing and algorithm development for detection and tracking (C/C++, Python, MATLAB)