Applied Research Scientist, AI Research

About the Role

Descript's Research team builds the models behind the product's most distinctive features: Video Regenerate and lipsync, video translation, zero-shot voice and roomtone cloning, and Studio Sound. We pick specific problems in the editing workflow and build specialized models for them.

This role is focused on multimodal understanding: training models to perceive edited media the way a human video editor does. Underlord, our AI editing agent, reasons about a project largely through a textual representation of it. Giving it direct perception of the media it's working on will let it judge its own output and reason about creative choices in an edit.

What You'll Do

• Multimodal understanding: Build vision-language systems that let Descript's agentic editing features reason over the visual and audio content of a project.
• Evaluation: Design benchmarks and evals that make editorial quality measurable and balance quality against cost and latency.
• Data: Build the datasets your work depends on, including synthetic data generation where real examples don't exist at scale.
• Training: Train specialized models from scratch or fine-tune existing foundation models.
• Shipping: Take models from prototype to production with the agent and engineering teams.
• Direction-setting: Identify the next research direction that should become a Descript feature. More senior candidates will own this directly; more junior candidates will grow into it.
• Publishing: Take work to academic venues if desired (supported, but not required).

Requirements

• Proven ability to design and implement deep learning algorithms, demonstrated by publications, open-source work, or shipped models.
• Strong programming skills and deep fluency in PyTorch.
• Track record of generating new ideas in machine learning and evaluating them quickly.
• Strong experimental judgment.
• Clear written and verbal communication.
• PhD or Master's in deep learning or a related field, or equivalent experience.
• At least one of the following:

- Lead or first author of an accepted publication in a top venue (CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR, or similar).
- Played a key role in shipping a production feature with deep learning as a core component.

• For Senior and Staff levels: track record of owning research direction and mentoring or technically leading researchers/engineers.

Breadth & Nice-to-Haves

• Vision-language models and multimodal understanding.
• Generative modeling for video, audio, or images.
• Post-training, fine-tuning, and RL on large foundation models.
• Building evaluation systems for generative or agentic outputs.
• Taking a research idea through to a shipped, production-facing feature.