Sr. Multimodal Model Training and Inference Optimization Engineer

Work with ByteDance’s Vision-Applied Research team to optimize training and inference for large multimodal generative AI models powering products like TikTok, CapCut, and Lemon8. You’ll improve large-model training pipelines, design distributed training strategies (data/model/pipeline parallelism and communication), and benchmark/profile deep learning workloads to remove bottlenecks and better utilize compute. Collaborate across research and engineering to scale efficient, high-performance model deployment.