Together AI
Research Engineer, Post-Training Inference
About this role
Join Together AI's Model Shaping team to build systems that enable developers to customize open-source foundation models with their own data. You'll work across the training and inference stacks, optimizing fine-tuning, reinforcement learning, and evaluation services while ensuring seamless deployment from post-training to production.
What you'll do
- Design and build systems for customizing open-source models at scale
- Integrate Model Shaping and Inference platforms for end-to-end post-training workflows
- Develop inference engine features and optimizations for RL training workloads
- Ensure platform stability, reliability, and 24/7 availability through on-call rotation
- Collaborate with product, research, and engineering teams on API design and integration
- Optimize inference performance for low-precision models and distributed GPU setups
What they're looking for
- Python or Go software engineering
- Inference engines (vLLM, SGLang, TensorRT-LLM)
- LLM fine-tuning methods and techniques
- Production ML systems design and deployment
- CUDA/Triton kernel development (preferred)
- Kubernetes for ML workload management (preferred)
- RL training optimization
- Multi-LoRA and distributed model serving
Benefits
- Competitive salary ($200,000–$290,000 base)
- Startup equity
- Health insurance
- Work on foundational open-source AI infrastructure
- Collaborate with leading ML researchers
- San Francisco location
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Together AI
Together AI builds GPU compute infrastructure and open-source model customization platforms for AI developers and enterprises. The company is hiring for infrastructure operations, ML systems engineering, go-to-market technology, customer success, and GPU research roles.
- Website
- together.ai
Likely interview questions
- Walk us through your experience deploying production ML services—what were the biggest challenges you faced around reliability and performance?
- Which inference engines have you worked with, and how did you optimize them for your specific use cases?