About the job
We're looking for a Senior Staff AI Software Engineer, Perception to own the technical vision and system architecture for how Digit perceives and stays safe as we take on these more dynamic, less constrained environments.
Responsibilities
Own the technical vision and system architecture for Perception that enables safety and efficiency, defining how Digit will perceive and reason about humans and hazards as it expands from structured work cells into open, dynamic environments
Analyze key architectural tradeoffs across sensing modalities, model design, redundancy, and compute budget, balancing safety, latency, and real-world robustness
Stay hands-on with development to prototype, validate, or build critical pieces yourself
Partner with functional safety and systems engineering to define the safety case strategy, identifying what architecture and evidence will be required
Support failure mode and hazard analysis at the system level, shaping which risks the architecture must design around
Evaluate emerging research and technology, and decide what's worth adopting into future robot architectures
Set standards for verification and validation of ML-based perception systems, especially for rare and safety-critical scenarios
Collaborate across navigation, controls, and hardware to make sure the perception architecture serves the whole robot
Act as a senior technical mentor and partner to engineers and technical leads across the organization, and represent the architecture to executive leadership, partners, and customers as needed
Qualifications
Minimum
8+ years in machine learning/perception for robotics, with a track record of defining system architecture
Master's or Ph.D. in Artificial Intelligence, Robotics, Computer Science, or a related field, with a strong foundation in machine learning, robotics, and intelligent systems
Demonstrated experience architecting ML/perception systems for safety-critical or safety-rated contexts — human detection, functional safety, automotive or robotics or a comparable regulated domain
Deep expertise in deep convolutional neural networks, multi-object tracking, data association, supervised learning, and pose estimation, with strong intuition for where these approaches break down in unpredictable, real-world conditions
Experience making system-level tradeoffs across sensing modalities, redundancy, and compute constraints on embedded/real-time hardware
A track record of setting technical direction at an organization level
Preferred
Publications in your field are a plus (CVPR, ICCV, RSS, ICRA preferred)
Direct experience supporting a safety case or certification submission
Experience architecting redundant or diverse sensing systems for safety-rated perception
Experience architecting perception systems for humanoid or legged robots specifically
Experience representing technical architecture to regulators, third-party certifiers, or safety auditors
Substantial experience with failure mode analysis (FMEA), hazard analysis, and safety case strategy for autonomous or robotic systems