Multimodal AI Model Optimization Research Engineer

What you'll do

• Apply sparsification, distillation, and quantization to research models.
• Own model optimization experiments, metrics, and benchmarking.
• Collaborate with researchers and engineers to deploy new ideas.

What they ask for

• Strong deep learning experience using PyTorch.
• Hands-on model optimization and compression experience, including distillation, pruning/sparsification, quantization, and mixed precision.
• Strong Python skills and understanding of inference performance and GPU/accelerator fundamentals.
• Experience with large models and datasets in cloud environments.

Nice to have

• Optimization experience with diffusion, video/audio generative, or large language models.
• Familiarity with TensorRT, ONNX Runtime, TVM, Triton, or XLA; or custom Triton/CUDA kernels and low-level performance tuning.

Tools

PyTorch, TensorRT, ONNX Runtime, TVM, Triton, XLA, CUDA, WebRTC

AI focus

Optimize and deploy multimodal AI models, including generative video/audio models and potentially diffusion models or large language models.