Location: Onsite-Austin, TX
Employment Type: Direct Hire,Full-Time
Job Title: Perception Lead, Staff Eng
Position Summary9 Mothers is seeking a Perception Lead to own the computer vision and machine learning stack that powers target detection, classification, and tracking in our counter-sUAS systems. The Perception Lead is responsible for the research, development, and deployment of the perception pipeline that enables EDDA to engage aerial threats in real time. This is a senior individual contributor position.
Essential Duties- Own the visual perception pipeline end-to-end, including detection, classification, and tracking of sUAS targets in real time.
- Design and train machine learning models that meet latency and accuracy requirements for edge deployment on Jetson-class hardware.
- Architect and maintain the dataset and simulation pipeline, including data collection, labeling, curation, augmentation, synthetic data generation, and closed-loop retraining based on field performance.
- Optimize inference performance on Jetson platforms, including model pruning, quantization, TensorRT integration, and custom kernel development as required.
- Establish the model evaluation framework and metrics used to assess perception performance under operational conditions.
- Deliver target state information (position, velocity, identity, uncertainty) to the controls subsystem.
- Collaborate with the hardware team on camera, optics, and sensor selection.
Requirements- 6+ years of professional experience in computer vision or machine learning, including production deployment on edge hardware.
- Demonstrated experience shipping a detection or tracking system under real-world latency, small-object, or adversarial constraints.
- Strong engineering skills beyond modeling, including experience with PyTorch or JAX through to deployed, optimized inference.
- Proficiency in Python for model development and C++ or Rust for deployed inference.
- U.S. citizenship and ability to pass a background check.
Preferred Qualifications- TensorRT and NVIDIA Jetson deployment experience at production scale.
- Multi-object tracking under challenging conditions.
- Multi-modal sensor fusion experience (EO/IR, radar, acoustic).
- Synthetic data generation and sim-to-real methodologies.
- Prior experience in defense or other safety-critical computer vision applications.
- Active security clearance, or eligibility to obtain one.
- A passion for building robots or engineering projects as a hobby.
Benefits- Meaningful Early Equity: You aren't just an employee; you are a foundational owner. Your contributions directly drive the value of your stake in the company.
- Direct Roadmap Influence: Forget the bureaucracy of big defense. You will have a seat at the table, directly shaping our product and technology trajectory from day one.
- Mission-Critical Work: We don't build for "what if." We build systems the Department of War actively needs to counter immediate, real-world threats.
- The Builder's Playground: Work in a brand-new lab fully optimized for rapid prototyping, equipped with NVIDIA Jetsons, high-end scopes, and 3-D printers.
- 100% Employer-Paid Premiums: We cover 100% of your medical, dental, and vision insurance premiums and cover 50% of healthcare premiums for your dependents.
- Unlimited PTO: We value results, not clock-watching. Take the time you need to stay sharp and recharge.
- Zero Red Tape: You report to the founders. You have the autonomy to make technical decisions that would take months of committee approval at a larger firm.
- Austin-Based Culture: Join an onsite team in Austin, TX, where we prioritize high-bandwidth collaboration and rapid field-testing.
- Relocation Assistance: We want the best talent in the room. If you aren't in Austin yet, we'll help you get here, to make your transition to the Silicon Hills seamless.
About the Interview- Application screen phone call (30 min)
- Live coding test with the Director of Engineering (45 min via MS Teams)
- Paid take-home consulting agreement ($500, ~4-8 hours)
- Onsite interview in Austin
- Offer
