Research Scientist

Zoom • Seattle, WA

6d • Hybrid

About The Position

As a Research Scientist on Zoom’s GenAI Science Team, you will play a key role in developing Zoom’s proprietary large language models (LLMs). The LLMs power AI Companion features such as meeting summaries, intelligent search, and agentic AI features. This position is focused on advancing model performance, building scalable AI systems, and driving product-oriented innovation. You can expect to work in a highly collaborative environment that values both scientific rigour and practical product impact. The GenAI Science Team is responsible for building Zoom’s in-house LLMs that support millions of users through AI-powered meeting intelligence features. The team’s mission is to create models that enable authentic, agentic AI tasks across the Zoom ecosystem.

Requirements

• Possess a master’s degree or PHD in Computer Science, Artificial Intelligence, Machine Learning, Distributed Systems, or a related field.
• Have publications in top-tier AI conferences (e.g., ICML, NeurIPS, ICLR, ACL, EMNLP, NAACL, CVPR, ICCV, AAAI).
• Have proficiency in Python, PyTorch and Hugging Face Transformers (required); Verl and Megatron.
• Have LLM post-training techniques such as RLHF, RLVR, reward modeling.
• Have experience in building evaluation systems and benchmarks for AI models, including human-aligned metrics and annotation quality management.
• Demonstrate advanced agent systems with function calls or tool usage.
• Be able to handle long-context modeling or retrieval-augmented generation.

Responsibilities

• Conducting research and development on large language models to support Zoom’s AI Companion and meeting intelligence features.
• Designing and implement model training pipelines, including data creation, filtering, and evaluation workflows.
• Collaborating with cross-functional teams of scientists, engineers, and product managers to translate research into production-ready solutions.
• Exploring and apply advanced post-training techniques such as RLHF, RLVR, and reward modeling.
• Developing evaluation systems and benchmarks to measure model quality, human alignment, and task performance.
• Contributing to the development of agentic AI systems capable of tool usage, function calling, and multi-modal understanding.

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level
• Number of Employees: 5,001-10,000 employees