About the job
Together AI is building the Inference Platform that brings the most advanced generative AI models to the world. Our platform powers multi-tenant serverless workloads and dedicated endpoints, enabling developers, enterprises, and researchers to harness the latest LLMs, multimodal models, image, audio, video, and speech models at scale.
Responsibilities
Build and optimize global and local request routing, ensuring low-latency load balancing across data centers and model engine pods.
Develop auto-scaling systems to dynamically allocate resources and meet strict SLOs across dozens of data centers.
Design systems for multi-tenant traffic shaping, tuning both resource allocation and request handling — including smart rate limiting and regulation — to ensure fairness and consistent experience across all users.
Engineer trade-offs between latency and throughput to serve diverse workloads efficiently.
Optimize prefix caching to reduce model compute and speed up responses.
Collaborate with ML researchers to bring new model architectures into production at scale.
Continuously profile and analyze system-level performance to identify bottlenecks and implement optimizations.
Qualifications
Minimum
5+ years of demonstrated experience building large-scale, fault-tolerant, distributed systems and API microservices.
Strong background in designing, analyzing, and improving efficiency, scalability, and stability of complex systems.
Excellent understanding of low-level OS concepts: multi-threading, memory management, networking, and storage performance.
Expert-level programming in one or more of: Rust, Go, Python, or TypeScript.
Bachelor’s or Master’s degree in Computer Science, Computer Engineering, or related field, or equivalent practical experience.
Preferred
Knowledge of modern LLMs and generative models and how they are served in production is a plus.
Experience working with the open source ecosystem around inference is highly valuable; familiarity with SGLang, vLLM, or NVIDIA Dynamo will be especially handy.
Experience with Kubernetes or container orchestration is a strong plus.
Familiarity with GPU software stacks (CUDA, Triton, NCCL) and HPC technologies (InfiniBand, NVLink, MPI) is a plus.