This role pays less than
91%
of similar roles. Most pay
$146,125–$231,150
— the shaded band above.
At the midpoint, this role pays about
$114k
versus about
$189k
for comparable roles.
Based on 240 similar postings.
Employer
About Booz Allen Hamilton
Booz Allen Hamilton is a management and technology consulting firm that provides analytics, digital, engineering, and cybersecurity solutions primarily to U.S. government agencies and commercial clients. Industry: Management & Technology Consulting
Booz Allen Hamilton currently has
802 open roles
on FindRole.
Listed pay typically runs
$86,800–$198,000
across 783 roles with salary data.
Computer Vision Software Engineer, Mid joins an engineering team developing advanced computer vision, multi-sensor fusion, and detection and tracking capabilities for remote sensing and GEOINT environments. The role involves designing, implementing, and optimizing CV algorithms and deep learning pipelines to solve problems involving target detection, characterization, and tracking across single and multi-sensor data. Key responsibilities include building training and deployment pipelines using Python or C++, integrating Kalman-family filters with modern workflows, and developing GPU-accelerated models using CUDA or RAPIDS. The position requires expertise in transformer-based architectures like DINO, CLIP, or SAM to improve image understanding. Candidates will work on microservice-based inference pipelines and containerized deployments for real-time environments. This role addresses technical challenges in national security missions by creating robust detection and track fusion algorithms for complex sensor data within constrained compute environments.
Develop and implement computer vision algorithms for target detection, characterization, and tracking across multi-sensor data.
Build and maintain model training, evaluation, and deployment pipelines using Python or C++.
Apply deep learning architectures, including transformer-based models like DINO, CLIP, and SAM, to image understanding problems.
Integrate classical estimation methods, such as Kalman filters, with modern deep learning workflows for real-time applications.
Develop GPU-accelerated models using CUDA, RAPIDS, or other vendor-provided inference runtimes.
Implement microservice-based inference pipelines and containerized model deployments.
Architect and implement novel single and multi-sensor platform detection and track fusion algorithms.
What we're looking for
Bachelor's degree in a STEM field.
Active TS/SCI clearance and willingness to take a polygraph exam.
3+ years of experience developing computer vision algorithms for detection, tracking, or sensor-based analytics in remote sensing or GEOINT environments.
1+ years of experience applying deep learning to CV problems using transformer-based, self-supervised, or contrastive learning architectures.
Experience building pipelines in Python or C++ for algorithm development, training, evaluation, or deployment.
Experience with GPU-accelerated workflows and programming, including CUDA or RAPIDS.
Experience with classical tracking and estimation methods, such as Kalman or extended Kalman filters.
Knowledge of modern software design patterns, including microservice design and Kubernetes orchestration.
Computer Vision
Deep Learning
Python
C++
CUDA
RAPIDS
TensorRT
Kubernetes
Docker
CI/CD
MLOps
ONNX Runtime
Triton Inference Server
multi‑sensor fusion
Agile
AWS GovCloud
Azure ML