Embedded Perception Engineer

ABOUT US: Collaborative autonomy is how self-tasking teams of machines will solve hard human problems, and HavocAI is an unquestioned leader in collaborative autonomy. We set the standard for autonomous surface vessels for a wide range of defense and commercial maritime missions. Success requires us to grow quickly, and we’re looking for teammates who are passionate about solving hard problems, about pushing the envelope, and about preventing conflict and saving lives. Ambition is welcome to apply within. ABOUT THE ROLE: As an Embedded Perception Engineer, you will own how our autonomous systems perceive and understand the world — in real time, in operational environments, and under demanding maritime conditions. This role goes beyond model training. You will architect, deploy, and optimize perception pipelines that run reliably on embedded edge hardware. You’ll work across vision models, sensor fusion, and high-performance inference to deliver robust situational awareness to autonomous surface vessels executing real-world missions. If you thrive in dynamic environments, enjoy squeezing maximum performance from embedded AI systems, and want to see your work field-tested and mission-proven, this role offers immediate and meaningful impact. KEY RESPONSIBILITIES AND REQUIREMENTS: PERCEPTION SYSTEM OWNERSHIP - Own end-to-end perception pipelines deployed to operational systems - Deliver high-reliability solutions aligned with mission requirements VISION MODEL DEVELOPMENT - Design, train, and optimize perception models for maritime environments - Develop object detection, classification, and tracking systems - Work with large vision models (ViT, CLIP, SAM, or similar) SENSOR FUSION & INTEGRATION - Fuse multi-modal sensor data (camera, radar, lidar) into unified perception outputs - Collaborate with hardware teams to select and integrate sensors - Develop and validate low-level sensor drivers for high-integrity data pipelines EMBEDDED DEPLOYMENT & OPTIMIZATION - Deploy models using NVIDIA Triton Inference Server - Optimize inference for latency and throughput on edge hardware - Apply quantization, pruning, and other optimization techniques CROSS-TEAM COLLABORATION - Integrate perception outputs into autonomy stacks alongside mission software and platform teams - Support troubleshooting of perception performance in operational environments FIELD PERFORMANCE & MONITORING - Define metrics and evaluation frameworks to monitor deployed system health - Investigate perception failures and drive resolution across teams END-USER ENGAGEMENT - Partner with operators and customers to translate field challenges into engineering solutions QUALIFICATIONS: - Bachelor’s or higher in Computer Science, Electrical Engineering, Robotics, Machine Learning, or related field - 3+ years developing and deploying computer vision or perception systems in C++ and/or Python on Linux - Experience with modern deep learning frameworks (PyTorch, TensorFlow) - Hands-on experience with large vision models (ViT, CLIP, SAM, etc.) - Experience implementing sensor fusion across heterogeneous modalities - Proficiency with NVIDIA Triton, TensorRT, ONNX Runtime, or equivalent inference tooling - Ability to quickly navigate and understand complex codebases - Strong systems-thinking mindset - U.S. Citizenship and eligibility for U.S. security clearance PREFERRED SKILLS: - Experience deploying perception systems on embedded platforms (Jetson, FPGA, similar) - Familiarity with maritime or aerial sensor environments - Experience optimizing models for edge inference - Knowledge of autonomy frameworks (ROS2, MOOS-IvP) - Experience with safety-critical or real-time systems - Familiarity with military systems and defense acquisition processes WHY THIS ROLE MATTERS You won’t just build models — you will build perception systems that operate in real-world, contested environments. Your work will directly influence the reliability, safety, and mission success of autonomous platforms depl