Computer Vision and Perception Engineer

ABOUT THE ROLE

Vanthal Technologies is seeking a Computer Vision and Perception Engineer to develop real-time perception systems for unmanned aircraft and collaborative robotic platforms.

This is a full-time, on-site position based at Vanthal’s operations in Annapolis, Maryland. The role includes model development, perception-software engineering, dataset evaluation, edge deployment, laboratory integration, simulation, and live aircraft testing.

The Computer Vision and Perception Engineer will develop systems for object detection, classification, tracking, visual localization, mapping, scene understanding, video processing, and perception-output generation.

The successful candidate must be capable of taking perception capabilities beyond research notebooks and deploying them as reliable, measurable software on resource-constrained edge computers.

This is not a research-only, data-analysis-only, web-development, or generic artificial-intelligence position. The role requires production software, physical-system integration, performance measurement, and operation under real-world visual and computational constraints.

RESPONSIBILITIES

• ■Develop real-time computer-vision and perception software
• ■Develop object-detection and object-classification systems
• ■Develop multi-object tracking systems
• ■Develop visual localization and visual-odometry capabilities
• ■Develop visual SLAM and mapping capabilities where appropriate
• ■Develop image registration, feature tracking, and motion-estimation methods
• ■Develop terrain, structure, vehicle, person, and environmental feature detection where appropriate
• ■Develop perception outputs for navigation, mapping, situational awareness, and human-supervised mission systems
• ■Develop methods for estimating object position, movement, confidence, and uncertainty
• ■Develop camera-calibration and sensor-calibration processes
• ■Develop video stabilization, enhancement, and preprocessing pipelines
• ■Improve performance under motion blur, vibration, compression, low light, glare, smoke, partial occlusion, and changing viewpoints
• ■Develop perception pipelines that operate under limited compute, power, storage, and bandwidth
• ■Optimize models and algorithms for edge-compute deployment
• ■Convert and deploy models using supported inference runtimes
• ■Measure latency, throughput, memory use, power use, accuracy, and reliability
• ■Develop confidence thresholds and failure-detection behavior
• ■Identify and reduce false positives and false negatives
• ■Develop dataset collection, curation, annotation, evaluation, and versioning processes
• ■Define representative training, validation, and test datasets
• ■Develop repeatable model-evaluation and regression-testing processes
• ■Analyze performance across environmental and operational conditions
• ■Integrate perception outputs with autonomy and mission-systems software
• ■Integrate cameras and perception pipelines with embedded edge computers
• ■Integrate timestamped imagery with aircraft telemetry and navigation data
• ■Support multi-sensor fusion where technically appropriate
• ■Build simulation, replay, and recorded-data test tools
• ■Develop automated tests for perception software and interfaces
• ■Analyze recorded video, logs, telemetry, and model outputs
• ■Diagnose model, software, camera, synchronization, calibration, and hardware-integration problems
• ■Support laboratory, field, and flight-test activity
• ■Maintain clear documentation for models, datasets, interfaces, calibration, evaluation, and deployment
• ■Participate in code review and software-quality practices
• ■Work directly with autonomy, embedded, mission-systems, networking, flight-test, and aircraft personnel
• ■Contribute to perception-system architecture and technical planning

REQUIRED QUALIFICATIONS

• ■At least four years of professional experience in computer vision, machine learning, robotics, perception software, imaging, or a closely related technical field
• ■At least two years of experience deploying computer-vision or machine-learning systems on physical hardware, robotic systems, vehicles, embedded computers, or edge-compute platforms
• ■Strong professional experience with Python
• ■Professional experience with modern C++
• ■Experience with deep-learning frameworks such as PyTorch or an equivalent framework
• ■Experience with OpenCV or comparable computer-vision libraries
• ■Experience developing object-detection, classification, tracking, localization, mapping, or comparable perception systems
• ■Experience training, evaluating, and deploying machine-learning models
• ■Experience developing repeatable model-evaluation methods
• ■Experience measuring precision, recall, false-positive rates, false-negative rates, latency, and computational performance
• ■Experience working with image and video datasets
• ■Experience with camera calibration and image-processing pipelines
• ■Experience deploying inference workloads on GPUs, accelerators, or edge-compute hardware
• ■Experience developing Linux-based software
• ■Experience integrating perception software with physical sensors or cameras
• ■Experience debugging issues spanning models, software, cameras, timing, hardware, and configuration
• ■Experience using Git and collaborative software-development practices
• ■Experience writing automated tests
• ■Strong understanding of probability, uncertainty, geometry, linear algebra, and coordinate transformations
• ■Strong understanding of software architecture and modular interface design
• ■Ability to translate operational requirements into measurable perception performance
• ■Ability to document datasets, models, assumptions, thresholds, limitations, and failure modes
• ■Ability to communicate technical results clearly
• ■Ability to work directly with aircraft, embedded hardware, and test personnel
• ■Willingness to participate in hands-on laboratory and field testing
• ■Reliable attendance and professional follow-through

PROGRAMMING LANGUAGE REQUIREMENTS

Python is required. Modern C++ is required.

MACHINE-LEARNING FRAMEWORK REQUIREMENTS

Professional experience with PyTorch or an equivalent deep-learning framework is required.

EDUCATION

A bachelor’s degree in computer science, computer engineering, electrical engineering, robotics, applied mathematics, physics, aerospace engineering, or a related technical field is preferred. Directly relevant computer-vision, perception, machine-learning, and software-engineering experience may be considered in place of a degree.

PREFERRED QUALIFICATIONS

• ■Six or more years of professional computer-vision, perception, machine-learning, or robotics experience
• ■Four or more years deploying perception systems on physical robotic or vehicle platforms
• ■Experience with real-time object detection and classification
• ■Experience with multi-object tracking
• ■Experience with visual odometry
• ■Experience with visual SLAM
• ■Experience with three-dimensional reconstruction
• ■Experience with structure from motion
• ■Experience with optical flow
• ■Experience with semantic or instance segmentation
• ■Experience with depth estimation
• ■Experience with camera pose estimation
• ■Experience with feature detection, matching, and geometric verification
• ■Experience with multi-camera systems
• ■Experience with electro-optical or infrared imaging systems
• ■Experience with sensor fusion
• ■Experience integrating cameras with inertial, navigation, or aircraft telemetry data
• ■Experience with geospatial coordinate systems and object geolocation
• ■Experience with PyTorch
• ■Experience with TensorRT
• ■Experience with ONNX or ONNX Runtime
• ■Experience with CUDA
• ■Experience with OpenCV
• ■Experience with NVIDIA Jetson or comparable edge-compute hardware
• ■Experience optimizing neural networks through quantization, pruning, compilation, or architecture selection
• ■Experience with GStreamer or comparable low-latency video frameworks
• ■Experience with ROS or ROS 2
• ■Experience with recorded-data replay and simulation systems
• ■Experience with distributed or multi-vehicle perception systems
• ■Experience with dataset versioning, annotation workflows, and experiment tracking
• ■Experience with active learning or hard-example mining
• ■Experience developing synthetic or augmented training data
• ■Experience evaluating models across environmental and operational conditions
• ■Experience with model monitoring and regression testing
• ■Experience with unmanned aircraft, ground robots, maritime robots, autonomous vehicles, or industrial robotics
• ■Experience supporting live robotic-system testing
• ■Experience in a startup or rapidly developing hardware organization
• ■Military UAS, aviation, imagery, robotics, or perception-systems experience

WORKING CONDITIONS

This position includes software development, laboratory integration, dataset evaluation, and field testing with physical unmanned-aircraft systems.

BENEFITS

• ■Three weeks of paid time off per year
• ■Health, dental, and vision insurance
• ■Paid maternity and paternity leave
• ■Eligibility for an annual performance bonus based on team or department success