# Lambert Leong, PhD | Lead AI/ML Scientist

Author: Lambert Leong
Original: https://www.lambertleong.com/

Please cite the original webpage linked above.

---

Aloha

# Lambert Leong, PhD

## Lead AI/ML Scientist

At Optum, I lead AI/ML projects spanning healthcare foundation models, clinical generative AI, and responsible AI. I combine hands-on model development and evaluation with ownership of technical direction and delivery, building on experience in computer vision and applied ML research.

### Areas of focus

- Foundation Models & Generative AI
- Computer Vision & Multimodal Learning
- Responsible AI, Safety & Clinical Validation
- Agentic Workflows & Decision Systems
- Applied ML Research and Technical Leadership

[Explore my site →](https://www.lambertleong.com/#explore)[More about me](https://www.lambertleong.com/#about)

![Lambert Leong smiling in an aloha shirt](https://www.lambertleong.com/assets/images/home/lambert_prof_2023_512.webp)

Lambert Leong, PhD

01 / Featured

## Featured posts.

[![](https://www.lambertleong.com/assets/images/essays/agent-auc/agentic_auc_cover2.png)](https://www.lambertleong.com/thoughts/agentic-auc)

thoughts · Dec 2025

## [Agents Under the Curve (AUC)](https://www.lambertleong.com/thoughts/agentic-auc)

Agentic AI systems are on the rise in healthcare, but most only output binary decisions. This post provides options to turn those decisions into continuous scores so we can still use AUC for fair model comparison.

 [Read more → Agents Under the Curve (AUC)](https://www.lambertleong.com/thoughts/agentic-auc)

[![](https://www.lambertleong.com/assets/images/essays/auc_idi_nri/idi_example.png)](https://www.lambertleong.com/thoughts/auc-idi-nri)

thoughts · May 2021

## [Area Under the Curve and Beyond with Integrated Discrimination Improvement and Net Reclassification](https://www.lambertleong.com/thoughts/auc-idi-nri)

Doctors and data scientist, here is an intro to IDI and NRI with a Python coded example. AUCs may not always be enough and this post allows you to go beyond the curve

 [Read more → Area Under the Curve and Beyond with Integrated Discrimination Improvement and Net Reclassification](https://www.lambertleong.com/thoughts/auc-idi-nri)

[![](https://www.lambertleong.com/assets/images/projects/13chip50_home_anim.webp)](https://www.lambertleong.com/projects/masters-thesis)

projects · Dec 2018

## [Designing 3D Multip Processor Compute Architectures (Masters Thesis)](https://www.lambertleong.com/projects/masters-thesis)

A Heuristic for Optimizing the Physical Layout and Network Topology of Integrated 3D Multi-chip Systems Under Temperature Constraints

 [Read more → Designing 3D Multip Processor Compute Architectures (Masters Thesis)](https://www.lambertleong.com/projects/masters-thesis)

02 / A little about me

## Curiosity is
the common thread.

I’m originally from Hawaii. I enjoy building things and exploring new trails. This site brings together my projects, thoughts, and personal interests.

Professionally, I’m a Lead AI/ML Scientist and team lead, with hands-on technical contributions and ownership of project outcomes. Healthcare is my strongest application domain. My background spans biology, computer science, and a PhD in Bioengineering. I care about useful systems and careful measurement.

[More about me ↗](https://www.lambertleong.com/about)[My professional background ↗](https://www.lambertleong.com/career)

03 / From around the site

## Recent thoughts & projects.

[![](https://www.lambertleong.com/assets/images/essays/system-1/routing-cover-green-part-1.png)](https://www.lambertleong.com/thoughts/system-1-decision-models)

thoughts · Nov 2026

## [When All You Have Are Decoders, Every Decision Looks Like Generation](https://www.lambertleong.com/thoughts/system-1-decision-models)

Agentic systems often use generative decoders for small routing decisions. This post proposes a decision layer—alongside, not instead of, generation—for faster, more measurable workflows.

 [Read more → When All You Have Are Decoders, Every Decision Looks Like Generation](https://www.lambertleong.com/thoughts/system-1-decision-models)

[![](https://www.lambertleong.com/assets/images/essays/agent-auc/agentic_auc_cover2.png)](https://www.lambertleong.com/thoughts/agentic-auc)

thoughts · Dec 2025

## [Agents Under the Curve (AUC)](https://www.lambertleong.com/thoughts/agentic-auc)

Agentic AI systems are on the rise in healthcare, but most only output binary decisions. This post provides options to turn those decisions into continuous scores so we can still use AUC for fair model comparison.

 [Read more → Agents Under the Curve (AUC)](https://www.lambertleong.com/thoughts/agentic-auc)

[![](https://www.lambertleong.com/assets/images/projects/pseudo_dxa/output3_home_anim.webp)](https://www.lambertleong.com/projects/pseudo-dxa-gen-ai)

projects · Dec 2024

## [AI Predicts Your Insides From Your Outsides With Pseudo-DXA](https://www.lambertleong.com/projects/pseudo-dxa-gen-ai)

A Quantitatively Accurate and Clinically Useful Generative Medical Imaging Model.

 [Read more → AI Predicts Your Insides From Your Outsides With Pseudo-DXA](https://www.lambertleong.com/projects/pseudo-dxa-gen-ai)

[![](https://www.lambertleong.com/assets/images/work/Optum-UHG/onsite-team1.jpeg)](https://www.lambertleong.com/experience/optum-ai)

experience · Sep 2024

## [Optum AI: Healthcare Foundation Models and Clinical GenAI Evaluation](https://www.lambertleong.com/experience/optum-ai)

I work at Optum AI as a Lead AI/ML Scientist, building health AI systems that span healthcare foundation models, structured EHR sequence modeling, clinical GenAI evaluation, model monitoring, and responsible AI review.

 [Read more → Optum AI: Healthcare Foundation Models and Clinical GenAI Evaluation](https://www.lambertleong.com/experience/optum-ai)

[![](https://www.lambertleong.com/assets/images/work/UW-postdoc/UW-imaging-reseach-lab-2024.jpg)](https://www.lambertleong.com/experience/university-of-washington-postdoc)

experience · Aug 2023

## [University of Washington OncoRad Research Core](https://www.lambertleong.com/experience/university-of-washington-postdoc)

I worked as a Postdoctoral Researcher at the University of Washington School of Medicine, building translational imaging, NLP, and computer vision tools for predictive biomarker research and clinical trial support.

 [Read more → University of Washington OncoRad Research Core](https://www.lambertleong.com/experience/university-of-washington-postdoc)

[![](https://www.lambertleong.com/assets/images/projects/tbdxa/DXA_image_types.png)](https://www.lambertleong.com/projects/final-dxa-nation)

projects · Jul 2023

## [Final DXA-nation](https://www.lambertleong.com/projects/final-dxa-nation)

AI and deep learning predict all-cause mortality from single and sequential body composition imaging, showcasing the power of longitudinal models in healthcare.

 [Read more → Final DXA-nation](https://www.lambertleong.com/projects/final-dxa-nation)

[![](https://www.lambertleong.com/assets/images/projects/photo_mosaic/web_gumwall_zoom_home_anim.webp)](https://www.lambertleong.com/projects/photo-mosaic)

projects · Jul 2022

## [Photo Mosaics with Nearest Neighbors: Machine Learning for Digital Art](https://www.lambertleong.com/projects/photo-mosaic)

We use machine learning to create digital art. K-nearest neighbors is used to build photo mosaics

 [Read more → Photo Mosaics with Nearest Neighbors: Machine Learning for Digital Art](https://www.lambertleong.com/projects/photo-mosaic)

[![](https://www.lambertleong.com/assets/images/projects/deep_learning_breast_cancer_risk/AI_breast_cancer_risk.png)](https://www.lambertleong.com/projects/deep-learning-breast-cancer-risk)

projects · Feb 2022

## [Artificial Intelligence Can See Breast Cancer Before it Happens](https://www.lambertleong.com/projects/deep-learning-breast-cancer-risk)

Deep Learning Predicts Interval and Screening-detected Cancer from Screening Mammograms, A Case-Case-Control Study in 6369 Women

 [Read more → Artificial Intelligence Can See Breast Cancer Before it Happens](https://www.lambertleong.com/projects/deep-learning-breast-cancer-risk)

[![](https://www.lambertleong.com/assets/images/projects/nature_commsmed_3cb_icad/heatmap_3cb_breast_lesions_home.webp)](https://www.lambertleong.com/projects/compositional-breast-cancer-ai)

projects · Nov 2021

## [Compositional Imaging with Artificial Intelligence to Improve Breast Cancer Detection](https://www.lambertleong.com/projects/compositional-breast-cancer-ai)

We use dual-energy X-ray compositional breast imaging with artificial intelligence and deep learning to improve cancer detection

 [Read more → Compositional Imaging with Artificial Intelligence to Improve Breast Cancer Detection](https://www.lambertleong.com/projects/compositional-breast-cancer-ai)
