Computer Vision Engineer

Netanya

Description

DRS RADA is a leading defense high-tech company specializing in radar systems development.

We are looking for a Computer Vision Engineer with a strong classical foundation who can also put the modern AI stack to work. The initial focus is the detection and tracking of small, dim targets in infrared imagery.

The role can grow into other areas, such as visual-inertial odometry, multi-sensor perception and more advanced AI pipelines. We are looking for someone who understands geometry and estimation from first principles and knows when a learned model is the right tool.


Key responsibilities


•         Perception pipelines: design and build end-to-end systems, from raw sensor data to real-time outputs.

•         Learning-based vision: develop and fine-tune models for detection, segmentation, tracking and related tasks, including hard cases such as small, low-contrast targets.

•         Classical CV and geometry: apply multi-view geometry, calibration, registration and feature-based methods where they are the stronger or complementary tool.

•         State estimation: build Kalman-family filters, multi-target tracking and visual or visual-inertial odometry.

•         Sensors: work across IR/thermal, visible and IMU data, including calibration, time sync and sensor-specific artifacts.

•         Modern AI: evaluate and integrate foundation models, vision-language models, self-supervised pretraining, synthetic data and multi-stage AI pipelines where they add value.

•         Data and evaluation: own collection plans, labeling, benchmarks and metrics tied to system performance.

•         Deployment: take models to production on edge and server hardware, optimizing for latency and reliability.

Requirements

Required qualifications


•         B.Sc. or M.Sc. in Electrical Engineering, Computer Science, Physics, Mathematics or a related field.

•         3 - 5 years of hands-on computer vision work, with systems shipped beyond the prototype stage.

•         Strong classical CV: projective geometry, homography, camera calibration, image registration, feature matching and robust estimation (RANSAC).

•         Hands-on deep learning in PyTorch: training, fine-tuning and debugging vision models.

•         Fluency with the modern AI stack: pretrained and foundation models, experiment tracking and model optimization.

•         Strong Python skills.

•         Strong math: linear algebra, probability, optimization and estimation theory.

•         Ability to take an open-ended problem from research to a working system, independently.


Nice to have


•         VIO and SLAM: filter-based (e.g. MSCKF) or optimization-based (e.g. VINS, factor graphs with GTSAM or Ceres).

•         IR/thermal imaging and small or dim target detection.

•         Foundation and vision-language models, open-vocabulary detection, and multi-stage or agentic AI pipelines.

•         Background in Kalman filters (KF, EKF, UKF) and tracking, used in real systems.

•         Multi-sensor fusion (EO/IR, IMU, GNSS, radar) and IMM or particle filters.

•         Edge deployment with TensorRT, CUDA or Jetson.

•         Synthetic data, simulation and self-supervised learning.

•         Publications or open-source contributions in vision, estimation or robotics.

Apply for this position