Applied Research Scientist, AI Research
$197k - $263k • Remote • San Francisco, CA or Remote, US
Posted 10d ago
Job Location
San Francisco, CA or Remote, US
Tech Stack
Remote Work Policy
Fully remote
Categories
Applied AI Engineer
About the job
Descript's Research team focuses on building specialized models for specific editing workflow problems, aiming to ship them as production features used by millions of creators. This role is centered on multimodal understanding, developing models that perceive edited media like a human video editor. The goal is to enhance Descript's AI editing agent, Underlord, by enabling it to directly perceive media, judge its own output, and reason about creative choices beyond just project structure. This is an open research problem with a unique dataset available.
Responsibilities
- Build vision-language systems for agentic editing features to reason over visual and audio content.
- Design benchmarks and evaluations for editorial quality, balancing quality, cost, and latency.
- Create datasets, including synthetic data generation, to support research work.
- Train specialized models from scratch or fine-tune existing foundation models.
- Collaborate with agent and engineering teams to take models from prototype to production.
- Identify and drive new research directions that can become Descript features.
- Publish work in academic venues if desired.
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 novel machine learning ideas and efficiently running experiments.
- Strong experimental judgment and ability to honestly assess idea viability.
- Clear written and verbal communication skills.
- PhD or Master's in deep learning or a related field, or equivalent experience.
- Either lead/first author of a publication in a top AI venue (CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR, or similar) OR played a key role in shipping a production deep learning feature.
- For senior/staff candidates: track record of owning research direction and experience mentoring or leading others.
Benefits
- Equity
- Benefits