Senior / Staff Research Engineer, Generative Ai Deployment

AI Summary

Job Description

As a Research Engineer focused on Generative AI Deployment, you will own the path from research to production for neural and generative models powering Waabi World. You will partner with research teams, making models production-ready: fast, memory-efficient, reproducible, and numerically trustworthy at scale.

Hard Requirements

• Experience reading and developing production-quality software at scale
• Shipped ML systems that other engineers and researchers depend on
• Taken models from a research codebase to an optimized production artifact
• Hands-on experience with TensorRT, torch.compile, torch.export, ONNX, post-training quantization, quantization-aware training, and pruning
• Handled models with custom CUDA operators or other operations that don't export cleanly
• Profile with tools such as Nsight Systems, Nsight Compute, and the PyTorch profiler
• Reason about memory bandwidth, kernel launch overhead, and host-device synchronization
• Write or modify CUDA or Triton kernels
• Read research code for modern architectures (vision transformers, diffusion and flow-matching models, vision foundation models)
• Know which accuracy trade-offs matter when lowering precision, compiling a model, or cutting its sampling steps
• Experience launching and debugging GPU jobs in the cloud (AWS, GCP, or equivalent)
• Job orchestration, containerized environments, and keeping CUDA, PyTorch, and TensorRT versions compatible

Soft Requirements

• Shipping Production Software
• Model Optimization and Deployment
• GPU Performance
• Modern ML Architectures
• Cloud Infrastructure