San FranciscoFull TimeEngineering
Remotely
pythonpytorchcudatritondeepspeedmegatron lmfsdp
Job Description
📋 Description Design, implement, and optimize core components of Together's large-scale training infra. Integrate new model architectures, validate training correctness, and optimize fine-tuning workloads. Profile distributed training to identify bottlenecks across compute, memory, and communication. Design and execute experiments to validate performance hypotheses and benchmark new approaches. Partner with researchers to productionize training methods and contribute to open-source releases. Rapidly enable support for newly released open-source foundation models on the Together platform.
Back to all jobs