Perception Engineer

Role:Perception Engineer
Location:San Francisco CA.
On site. Full time.

About the job:
Advantra is partnering with an early-stage defense technology startup in San Francisco to hire a Perception Engineer to build and deploy vision and sensor-fusion systems for edge-deployed laser defense platforms.

About the role:
You will own the end-to-end perception stack integrating multi-modal sensors developing real-time detection models and deploying ML and CV pipelines on edge hardware. You will collaborate with firmware hardware and controls teams to ensure high-throughput reliable perception for laser defense systems.

Responsibilities:

1. Design train validate and fine-tune machine learning and deep learning models for detection classification and segmentation.
2. Integrate and fuse data from RGB thermal LiDAR/ToF IMU and encoder sensors to generate robust real-time regions of interest.
3. Implement advanced computer vision algorithms for low SNR detection deconvolution and motion isolation.
4. Collaborate with hardware teams to integrate sensors over GigE Vision USB3 Vision CAN SPI MIPI and I C.
5. Develop and debug embedded firmware in C/C or Rust for microcontrollers and FPGAs within RTOS environments.
6. Build scalable data ingestion labeling augmentation and storage pipelines for simulated and field data.
7. Optimize inference frameworks for GPU/FPGA edge deployment.
   Develop dashboards and telemetry for drift analysis hardware monitoring and automated retraining triggers.
8. Author technical documentation and mentor junior engineers on sensor-fusion and embedded firmware best practices.
   
   

Core requirements:

1. 3 to 6 years in computer vision sensor-fusion or robotics perception roles.
2. Proficient in C and Python.
3. Hands-on experience with ML frameworks (TensorFlow PyTorch) and real-time inference engines (TensorRT OpenVINO).
4. Experience integrating and calibrating multi-modal sensors (cameras thermal imagers LiDAR/ToF IMUs).
5. Firmware development experience on microcontrollers (STM32 NXP TI) and FPGAs using VHDL/Verilog in RTOS environments.
6. Familiarity with Docker and CI/CD pipelines for ML workloads.
7. BS or MS in Computer Science Robotics Electrical Engineering Mathematics or related field.
8. Strong problem-solving communication and cross-functional collaboration skills.
9. Ability to deploy ML models and CV algorithms on edge hardware with high throughput.
10. Must be able to obtain a security clearance.
    
    

Nice to have:

1. Experience with Edge-AI optimization (quantization pruning).
2. FPGA or embedded GPU platform experience.
3. Background in defense or safety-critical systems.
4. Familiarity with cybersecurity guidelines and secure coding practices.

How to apply:
Email with links to shipped work and a short note on what you owned. Company name shared after the intro.

About Us:
Advantra-Upstart Crew is a search program inside Advantra Consulting. We partner with early stage and high growth startups to hire the top 2% in Tech and GTM. We run end to end searches from single role headhunts to full team build outs using domain experts and a vetted network to deliver tight shortlists. We stay close to founders and candidates so the relationship lasts beyond the first hire.