> Markdown version of [/jobs/ext/4488312-ai-engineer](https://www.wearedevelopers.com/jobs/ext/4488312-ai-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# AI Engineer

- **Company:** Zoom Video Communications, Inc.
- **Location:** Seattle, WA, United States (Remote available)
- **Experience:** Starter
- **Salary:** $137,700.0 - $275,400.0
- **Contract:** Internship / Graduate position
- **Skills:** Artificial Intelligence, C++ (Programming Language), Nvidia CUDA, Computer Programming, Computer Engineering, Software Debugging, Python (Programming Language), Linux Kernel, Machine Learning, Open Source Technology, Remote Direct Memory Access, Software Engineering, Graphics Processing Unit (GPU), Pytorch, Large Language Models, Information Technology, ONNX (Open Neural Network Exchange) Format, SGLang, Machine Learning Operations, TensorRT, VLLM, Model Inference, CUTLASS
- **Published:** October 31, 2026
- **Apply:** https://dejobs.org/x/x/3E0C3873A6BF4A92838C7E9E95CA5BA8/job/

## About the Role

* A degree in Computer Science, Computer Engineering or a related field (MS or PhD preferred), or equivalent practical experience
 * 0-3 years of software engineering experience; internships, research and open-source work in ML systems count
 * Hands-on experience with at least one inference framework (vLLM, SGLang, TensorRT-LLM or ONNX Runtime), at the level of serving and modifying it, not only exporting models
 * Working knowledge of how LLM inference performs on GPUs: memory bandwidth, KV cache, batching, and communication across GPUs
 * Experience profiling and debugging GPU performance, with evidence of an optimisation you measured and shipped
 * Depth in at least one of: speculative decoding, continuous batching, KV cache design, quantisation, distributed or multi-GPU serving
 * Strong programming skills in Python and C++, and comfort reading and changing large framework codebases

Preferred

 * Merged contributions to vLLM, SGLang, TensorRT-LLM, PyTorch or similar projects
 * CUDA, Triton or CUTLASS kernel development
 * Experience serving MoE or 100B+ parameter models, or multi-node inference (NCCL, RDMA/RoCE)
 * Graduate coursework or research in ML systems, compilers or high-performance computing

## Description

You will build and optimize the inference systems that serve Zoom's AI models in production. Working alongside senior inference engineers, research and platform teams, you will profile serving workloads, find the bottlenecks that limit latency, throughput and cost, and ship the fixes. You will work close to the metal: inside inference frameworks, GPU profiles and serving infrastructure., * Build, tune and maintain components of Zoom's inference serving stack for transformer and multimodal models, including large and Mixture-of-Experts (MoE) architectures

 * Implement and evaluate inference optimisations such as speculative decoding, continuous batching, KV cache management, parallelism layouts and quantisation (INT4/INT8/FP8)
 * Extend and customise inference frameworks (vLLM, SGLang, TensorRT-LLM or equivalent) to support new models and Zoom's production requirements
 * Profile GPU workloads with Nsight Systems or equivalent tools, root-cause performance regressions, and validate improvements with reproducible benchmarks
 * Build and improve internal tooling for benchmarking, performance modelling and serving configuration
 * Partner with researchers to bring new model architectures into production efficiently
 * Contribute to on-call and production support for the services you build, with guidance from senior engineers