About the job
About the Role
Descript is at the forefront of transforming traditional editing practices by seamlessly integrating AI technologies throughout the content creation journey—encompassing recording, transcription, editing, and sharing. Our goal is to simplify and accelerate the production of high-quality content.
We are seeking an Engineering Manager to spearhead our Agent team, dedicated to developing the premier Agentic Video Editor, Underlord. This dynamic field presents unique technical challenges in delivering high-quality, cost-effective Agents to navigate the multimodal complexities of video content. You will oversee a team of 5-6 software engineers, report directly to the Head of AI Engineering, and collaborate closely with various teams across engineering, product development, design, and AI research.
Agent-driven editing is a key focus for our company, and this role demands exceptional product insight and strategic partnership skills as the team innovates high-quality experiences in Agentic Video Editing.
About the Team:
Since launching in Open Beta in July 2025, Descript's Agent team has been tirelessly enhancing the user experience for chat-based editing. We are continually expanding our agent's capabilities and refining our infrastructure to keep pace with advancements in AI models. The Agent product engineering team maintains a close partnership with our AI Research team to fulfill our mission of making video editing accessible to everyone. Your management peers and team members are based in San Francisco, New York, and distributed across the United States and Canada.
This position is ideal for you if you are passionate about advancing in the following areas:
- Creating an unparalleled product experience for agentic video editing interactions (e.g., optimizing prompt templates and key user workflows, addressing user feedback, and implementing core product features like chat history).
- Leading rapid, data-driven experiments to enhance the quality of the video editing agent (e.g., conducting model/routing experiments, pricing tests, and investing in core agent functionalities).
- Enhancing the Agent Quality flywheel to prioritize our development roadmap by utilizing user data and feedback to improve evaluation and product analysis processes.
- Collaborating with AI Research to outline the roadmap for innovative Agent enhancements (e.g., exploring novel context management strategies and developing new models).
What You'll Do:
Your primary focus in 2026 will be to elevate quality as we transition from beta to General Availability (GA) and deliver significant impact for the business.

