> Markdown version of [/jobs/ext/4613132-lead-ml-scientist](https://www.wearedevelopers.com/jobs/ext/4613132-lead-ml-scientist). 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).

---

# Lead ML Scientist

- **Company:** Adobe Inc.
- **Location:** San Francisco, CA, United States
- **Experience:** Expert
- **Contract:** Permanent contract
- **Skills:** Machine Learning, Tensorflow, Reinforcement Learning, Pytorch, Large Language Models, Adobe, Information Technology, GPT
- **Published:** November 7, 2026
- **Apply:** https://www.juju.com/job/28_25303_39a764e4a60410b9245616f3fae97d55

## About the Role

* Master's degree or PhD in computer science, statistics, engineering, or other relevant subject area highly preferred
 * 8+ years of relevant experience required

## Description

* Research and implement ranking ML models using Transformers, GPT, reinforcement learning, diffusion models, TensorFlow, and PyTorch
 * Build a relevance platform to rank billions of images, documents, videos, and other assets in real-time.
 * Fine-tune LLMs and Diffusion models to build state-of-the-art ranking models
 * Conduct innovative research on novel LLM architectures and training methods, focusing on creative applications
 * Leverage the user's intent, context, deep content understanding, and behavioral data as ranking signals.
 * Design whole-page relevance by aggregating diverse content and ranking based on a query to asset-type affinity.
 * Process multimodal (text, image, video) intent and contextually assist the user by recommending ingredients and generating rich artifacts to enable a streamlined user journey.
 * Align text and visual modalities by fine-tuning CLIP for the creative domain
 * Unlock assistive experience in creative products through the adoption of foundation models
 * Develop robust methods to evaluate LLM performance, ensuring ethical AI practices that reflect Adobe's standards