Teams at Veeda AI

Design, deploy, and maintain large-scale bare-metal GPU clusters for distributed training. Tune Slurm schedulers and Kubernetes integrations, optimize storage and high-speed networking, and build observability pipelines to ensure reliable, high-throughput GPU workloads for distributed deep learning.

Work on product-driven research to develop multimodal generative world models for image, video, and 3D. Design and train foundation models, implement post-training alignment (RLHF, DPO, reward models), and build large-scale data curation pipelines for sourcing, filtering, and labeling training data.

Design, optimize, and maintain large-scale distributed multi-GPU training systems. Diagnose numerical precision and stability issues (FP16/BF16/FP8), detect and recover from hardware/software faults, profile performance bottlenecks, and build developer tooling for resilient checkpointing and rapid fault recovery to maximize researcher productivity and hardware utilization.