Veeda AI
Member of Technical Staff - ML Performance
Toronto +3 · Hybrid · Staff · Full-time · Posted Aug 19, 2026
- Location
- Toronto
Zürich
Seattle
California - Education
- Bachelor's degree or equivalent hands-on experience in Computer Science, Computer Engineering, or a related technical field.
About the job
Optimizes distributed training and inference performance, numerical stability, kernels, communication, and fault recovery for multimodal video world models.
What you'll do
- Improve multi-node training throughput and parallelism configuration.
- Profile and optimize inference latency, GPU utilization, precision, and kernels.
- Tune communication and build diagnostics and checkpointing for training failures.
What they ask for
- Bachelor's degree or equivalent hands-on experience in Computer Science, Computer Engineering, or a related technical field.
- Deep hands-on PyTorch experience and experience with at least one large-scale parallelism stack on real multi-node jobs.
- Fluency in Python and C++/CUDA.
- Experience profiling live training and inference runs with Nsight Systems or the PyTorch profiler; expertise in at least one of low-precision numerics, kernel authoring, or large-run fault diagnosis.
Nice to have
- Experience optimizing workloads across very large GPU clusters.
- Experience writing Triton, CUTLASS, or CuTe-DSL kernels, or contributing to open-source kernel libraries.
Tools
PyTorch, FSDP2, Megatron-Core, TorchTitan, DeepSpeed, Python, C++/CUDA, Nsight Systems
AI focus
Performance engineering for multimodal foundation world models for Physical AI, including video model training and inference.
Model-written summary · the original posting is authoritative