Research Engineer, Post-Training Inference

Research Engineer role on the Model Shaping team at Together AI, focusing on building a platform for customizing open-source models with user data. Responsibilities include developing fine-tuning, RL, and evaluation services, integrating training and inference stacks, optimizing inference engines for RL workloads, and ensuring platform reliability through on-call rotation.

2+ years building and deploying ML services in production. Hands-on experience with inference engines (SGLang, vLLM, TensorRT-LLM). Familiar with fine-tuning LLMs. Strong software engineering in Python or Go. Experience with low-precision models, Multi-LoRA, distributed models, RL workload optimization, CUDA/Triton kernels, large-scale systems, open-source ML contributions, and Kubernetes.