1 day ago
Base Salary
$138k - $275k/yr
Responsibilities
- Build, tune, and maintain Zoom’s inference serving stack for transformer and multimodal models, including large and Mixture-of-Experts architectures.
- Implement and evaluate inference optimizations such as speculative decoding, continuous batching, KV cache management, parallelism layouts, and INT4, INT8, and FP8 quantization.
- Extend and customize vLLM, SGLang, TensorRT-LLM, or equivalent inference frameworks for new models and production requirements.
- Profile GPU workloads, identify performance bottlenecks and regressions, and validate improvements with reproducible benchmarks.
- Build internal tooling for benchmarking, performance modeling, and serving configuration.
- Partner with research teams to bring new model architectures into production.
- Contribute to on-call and production support for the services developed.
Requirements
- Bachelor’s degree in Computer Science, Computer Engineering, or a related field, or equivalent practical experience; an MS or PhD is preferred.
- 0–3 years of software engineering experience, with internships, research, and open-source work in ML systems counted as relevant experience.
- Hands-on experience serving and modifying at least one inference framework, such as vLLM, SGLang, TensorRT-LLM, or ONNX Runtime.
- Working knowledge of LLM inference on GPUs, including memory bandwidth, KV caches, batching, and multi-GPU communication.
- Experience profiling and debugging GPU performance with evidence of a measured and shipped optimization.
- Depth in at least one area such as speculative decoding, continuous batching, KV cache design, quantization, distributed serving, or multi-GPU serving.
- Strong programming skills in Python and C++, with the ability to read and modify large framework codebases.
- Preferred qualifications include contributions to vLLM, SGLang, TensorRT-LLM, PyTorch, or similar projects; CUDA, Triton, or CUTLASS kernel development; experience with MoE or 100B+ parameter models or multi-node inference; and graduate coursework or research in ML systems, compilers, or high-performance computing.
Benefits
- Structured hybrid work approach centered around office and remote work environments.
- Benefits and perks supporting physical, mental, emotional, and financial health, work-life balance, and community involvement.
- Applications are accepted for at least five days, with an anticipated closing date of November 13, 2026.
Categories
About Zoom
Zoom builds a cloud platform for video meetings, team chat, webinars, phone, rooms, and contact center used by businesses, schools, and public-sector organizations. It sells these collaboration and communications services as SaaS subscriptions with add-ons for events, telephony, and enterprise features, and offers developer APIs/SDKs. Founded in 2013 and headquartered in San Jose, California, Zoom Video Communications is a public company traded on Nasdaq.