Role description
Veeda AIABOUT US
Veeda AI is building the next generation of multimodal foundation world models for Physical AI. We're a small, fast-moving team of engineers and researchers from leading AI labs, tackling some of the most challenging problems at the intersection of AI, robotics, and embodied intelligence. If you're excited about pushing the boundaries of what's possible with Physical AI, you'll have the opportunity to make an outsized impact from day one.
RESPONSIBILITIES
Own one scoped project in a single area — ML Performance, AI Infrastructure, ML Operations, Data, Robotics, Simulation, or World Models — with a mentor and a deliverable agreed in week one.
Representative Projects: Profile a step-time regression in a Megatron-Core run, add a task to an Isaac Lab environment suite, build a deduplication pass in Ray Data, or measure drift in a world-model rollout.
Real Systems, Not a Sandbox: Work in the same repositories, on the same clusters, and against the same data as the rest of the technical staff, not a parallel toy version.
A Deliverable That Outlives You: Ship one artifact the team keeps using — a dataset, a benchmark, an evaluation, or a tool — and write up what you found and what to try next.
REQUIREMENTS
You have a Bachelor's degree or equivalent hands-on experience in Computer Science, Engineering, or a related technical field, or you are currently working toward one.
You have real depth in machine learning fundamentals (optimization, generalization, architectures) or in systems fundamentals (operating systems, networking, parallel computing), plus the intuition to know when theory breaks down.
You program fluently in Python and have built and debugged a non-trivial system end to end, on your own.
You can name which of our seven areas you want to work in and sketch, concretely, what you would try to build or measure there.
NICE TO HAVE
You have built or maintained scalable data pipelines, training infrastructure, or simulation tooling.
You have published or contributed to research on generative models for image, video, or 3D content, or on robot learning.
You have hands-on experience with GPU kernels (CUDA, Triton), a physics engine (MuJoCo, Newton), or real robot hardware.
You have contributed to an open-source project that other people use.
You have self-directed projects, research competitions, or competitive programming results that show how quickly you pick things up.
You can work hybrid from one of our offices — Toronto, Mountain View, Zürich, or Singapore — for the duration of the internship.