Browse All Jobs
Job Description
Bridgewater Associates is seeking a Sr. Machine Learning Engineer (PyTorch) to join its AIA Labs, an in-house venture focused on using AI to generate returns in markets. The ideal candidate will build, optimize, and deploy ML models at scale, working closely with researchers to implement state-of-the-art models and accelerate training and inference. This role requires strong PyTorch expertise and a deep understanding of model optimization.

Role involves:
  • Developing and refining deep learning models in PyTorch.
  • Optimizing models for deployment using tools like TorchScript, ONNX, TensorRT, Triton, and torch.compile.
  • Identifying and resolving bottlenecks using torch.profiler, nvprof, Nsight, and memory profiling.
  • Building custom PyTorch modules and implement efficient backward passes when necessary.
  • Scaling models for distributed training and inference.
  • Translating research insights into robust, high-performance ML systems.

Requirements:
  • 5+ years’ experience in a technically rigorous environment with modern full stack tech stacks.
  • Strong PyTorch expertise (torch.nn, torch.autograd, torch.compile, torch.fx).
  • Solid understanding of model architectures (e.g., Transformers, CNNs) and deep learning fundamentals.
  • Experience with model optimization (e.g., mixed precision, checkpointing, memory-efficient execution).
  • Ability to profile and debug PyTorch performance (e.g., CUDA memory issues, kernel execution time).
  • Strong engineering background (Python, structured software development, scalable ML workflows).

Role offers:
  • Health insurance with 100% premium covered.
  • 401(k) plan with generous employer match.
  • Paid time off, including fully paid parental leave and a competitive PTO package.
  • Workplace flexibility and access to back up childcare.
  • Financial assistance for family building support, including adoption and egg freezing.
  • Workplace wellness, including on-site gyms, free meals and healthy snacks, and meditation rooms.
Apply Manually