Job Specifications
Job Description
AI Research Scientist II Intern, Physical AI
Location: Woburn, MA
Duration: Summer 2026 (3–6 months; 6 months preferred)
Department: Physical AI, AI and Data Innovation
The Opportunity
Chewy Robotics is seeking a PhD Research Intern to help advance fleet intelligence for heterogeneous humanoid and mobile robots—algorithms that decide which robot does what, when, and where under real-world constraints (e.g., congestion, priorities, uncertainty, safety), based out of our Robotics Lab in Woburn, MA.
Your work will contribute to Chewy’s fleet orchestration layer (RMS), which couples decision-making at scale, shared spatial understanding, and safety/resilience for multi-robot operation in dynamic, human-occupied environments. You will be expected to run simulation-first experiments and drive evaluation using clear metrics (e.g., throughput, coordination latency / replanning responsiveness, robustness, safety-event rate, localization/estimation error, map consistency, world-model update latency) and, where appropriate, validate results in lab experiments. Your findings are expected to result in publishable contributions in peer-reviewed conferences or journals.
Example Project Areas May Include
Shared spatial understanding (C-SLAM / perception / estimation): collaborative and active SLAM, semantic 3D mapping, multi-modal sensor fusion, map consistency and drift mitigation, fleetwide world-model maintenance.
Decision-making at scale: optimization-based task allocation and scheduling (LSTA), learning-assisted planning, constraint-aware dispatch under uncertainty.
Multi-robot coordination: multi-agent path finding (MAPF) and routing, congestion-aware replanning, fleet-level safety supervision and fail-safe behaviors.
What You’ll Do
Conduct end-to-end research (problem formulation, baselines, implementation, evaluation) under the direction of Chewy Research Scientists, with an emphasis on rigor and reproducibility.
Develop and implement algorithms in one primary area (e.g., C-SLAM/estimation, task allocation/scheduling, MAPF/routing, fleet safety supervision, or learning-assisted planning) and integrate outputs into a broader fleet-orchestration workflow when appropriate.
Build and instrument simulation-first experiments; define metrics, conduct ablations, and analyze results with publication-quality methodology.
Develop experimental protocols, datasets/benchmarks, and reusable evaluation harnesses with strong reproducibility standards.
Work multi-functionally with Chewy’s AI, robotics, and systems engineering teams to accelerate the deployment path from research to lab validation.
Prepare manuscripts, technical reports, and internal presentations which detail research progress.
What You’ll Need
Currently enrolled in PhD program in Robotics, Electrical Engineering, Mechanical Engineering, Computer Science, Applied Mathematics, or related field, with a focus on perception/estimation, controls, optimization, multi-agent systems, robotics, or AI.
Deep expertise in at least one of the following (with working familiarity across adjacent areas)
Multi-agent estimation/mapping (e.g., C-SLAM, semantic mapping, sensor fusion)
Multi-agent task allocation / fleet scheduling (LSTA)
Motion planning / MAPF / routing
Control + state estimation (e.g., MPC, filtering)
ML for decision-making (RL/IL/model-based)
Strong programming skills in at least one of Python, C/C++, ROS/ROS2, or Julia, and comfort working in robotics simulation environments (e.g., MuJoCo, Isaac Sim, PyBullet, Gazebo).
Strong experimental or analytical reasoning skills with demonstrated ability to develop, validate, and iterate on algorithms using data.
Proven publication record or clear evidence of research capability (e.g., top-tier venues such as NeurIPS, ICLR, ICML, ICRA, IROS, CDC).
Excellent written communication skills and motivation to publish scientific results.
Bonus
Experience with heterogeneous robot fleets and capability-aware sensing/scheduling/routing.
Familiarity with optimization modeling tools and solvers (e.g., CasADi, OR-Tools, and LP/QP/MIP/NLP formulations).
Experience with safety-critical robotics concepts (constraints, safety envelopes, monitoring, fallbacks) or high-availability systems.
Prior experience with hardware-in-the-loop testing or real-robot experimentation.
The hourly rate for this role is $50.00.
The specific salary offered to a candidate may be influenced by a variety of factors including but not limited to the candidate's relevant experience, education, and work location. In addition, this position is eligible for 401k and a new hire and annual equity grant. C08+ positions may also be eligible for annual bonus.
We offer different types of insurance and benefits, such as medical/Rx, vision, dental, life, disability, hospital indemnity, critical illness, and accident. We offer parental leave, family services benefits, backup dependent care, flexible spending accounts, telemedicine, pet adoption r