SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Archer Technologies is building an all-electric vertical takeoff and landing (eVTOL) aircraft designed to carry four passengers with minimal noise, advancing sustainable air mobility. As Staff Engineer, AI Autonomy, you will lead the development and deployment of AI systems that enable Archer's aircraft to perceive, reason, and navigate autonomously in the real world.
You will own the full lifecycle of AI autonomy systems, from research through production deployment. Key responsibilities include: developing and integrating multi-model systems including vision-language models (VLN), vision-language-action models (VLA), perception systems, and language models; designing comprehensive benchmarks and measuring success/failure modes against internal baselines and published results; fine-tuning and adapting foundation models while clearly weighing tradeoffs; validating models through data-driven testing in simulation, on hardware, and in actual flight tests in close partnership with Safety and Simulation Engineers; and taking models from prototype to deployed, monitored components running on edge hardware under real latency and power constraints.
You will work at the intersection of cutting-edge AI research and safety-critical aerospace systems, where your models directly impact aircraft behavior and safety. The role requires deep technical expertise in foundation models, transformers, and multimodal AI, combined with practical experience shipping ML systems to production hardware with strict performance requirements.
Required qualifications: 8+ years of experience in machine learning, AI, robotics, or related fields; M.S., Ph.D., or equivalent in Computer Science, Robotics, or related discipline (exceptional B.S. candidates considered); hands-on experience building, fine-tuning, and integrating vision-language or VLA models; solid understanding of transformer architectures and multimodal foundation models, including practical mechanics of fine-tuning pre-trained checkpoints to new domains.
