Ke Ni

Ke Ni

he/him

Applied Scientist at Amazon. The Joint CMU–Pitt Ph.D. Program in Computational Biology (CPCB). I work on multimodal models, LLMs, computer vision, and AI for science.

About me

TL;DR: I'm an Applied Scientist II at Amazon. I graduated from CPCB, the Joint CMU–Pitt Ph.D. program in Computational Biology. I was accepted by Carnegie Mellon University's School of Computer Science Ph.D. program (2019–2020) and continued at the University of Pittsburgh School of Medicine (2021–2025), cross-registered at both universities throughout the program. During 2020–2021, I was based in Shanghai through the COVID-19 pandemic, where I co-founded two startups, serving as research lead and machine learning engineer. Before that, I earned my B.S. and M.S. in the Department of Computer Science (College of Engineering) at UC Santa Barbara, advised by William Wang and the UCSB NLP group. Today, I am deeply grateful to Amazon for the opportunity to work on Responsible AI (safety).
More about me

I’m an Applied Scientist at Amazon, where I work on vision-language models, responsible AI, and interpretability research: post-training MLLMs/VLMs, mechanistic interpretability of embedding models, and applying generative AI to improve data efficiency and robustness. Since 2024, I have also been building agentic systems as a self-funded personal effort, starting from LLM-in-a-loop multi-agent designs, including a tool-use code agent, and progressing toward FARS.

I completed my Ph.D. in the Joint Carnegie Mellon University–University of Pittsburgh Program in Computational Biology (CPCB), defending my thesis in November 2025. The program spans Carnegie Mellon’s School of Computer Science and the University of Pittsburgh’s School of Medicine. During my doctoral research I collaborated with the Xing lab (University of Pittsburgh) and the Xu lab (Carnegie Mellon University), building deep-learning frameworks for live-cell imaging analysis (LivecellX and LivecellAction), single-cell genomics analysis, and cryo-EM image analysis. I am grateful to my collaborators for their generous partnership on our shared projects, including Xueying Zhan (then a postdoc in the Xu lab), Yuhao Chen (now an Investigator at Westlake University), and Yanshuo Chen (now a Ph.D. candidate at the University of Maryland). I also thank Dr. Xiaojie Qiu (now at Stanford University) for introducing me to single-cell genomics and RNA velocity research during my early Ph.D. years.

Earlier, I completed my B.S. and M.S. in Computer Science at the University of California, Santa Barbara, where Prof. William Yang Wang introduced me to natural language processing research; I deeply appreciate his mentorship. Before moving into research, I was a largely self-taught competitive programmer, taking part in the Informatics Olympiad (high school) and ICPC (college). In high school I learned through Coursera and was mentored by Dr. Rong Ma (Shanghai Jiao Tong University), who taught me to cultivate self-evolving learning skills rather than to memorize algorithms, a habit that has shaped how I approach problems ever since.

Throughout all of my education, I am deeply thankful for my family’s unwavering support.

Research interests: Multimodal Models (MLLMs/VLMs); Large Language Models; Mechanistic Interpretability; Reinforcement Learning & Post-training; Agentic Systems; Computer Vision; AI for Science and Single-cell Genomics.

News

Work from my doctoral years is marked CPCB

[2026] Together with my colleagues on the Amazon Science team, we submitted multiple papers to top ML/NLP/CV venues, stay tuned!

[2026] In April 2026, I joined Amazon as an Applied Scientist II, working on vision-language models, responsible AI, and interpretability research.

[2026] After several years of journal publications, I have shifted my focus back mainly to AI/ML conferences (EMNLP, AAAI, ECCV, and beyond). I remain deeply interested in interdisciplinary research, especially lab-in-the-loop agentic systems and lab automation (embodied intelligence for science discovery acceleration).

[2026] Score-Based Matching with Target Guidance for Cryo-EM Denoising was accepted to ECCV 2026. CPCB

[2025] I defended my Ph.D. thesis (November 2025). CPCB

[2025] LivecellX (in review, Nature Communications) and LivecellAction (in preparation): deep-learning frameworks for single-cell imaging analysis. CPCB

[2025] GraphVelo was published in Nature Communications. CPCB

[2024] Transiently Increased Coordination in Gene Regulation During Cell Phenotypic Transitions was published in PRX Life. CPCB

[2023] I presented at CytoData 2023 (Allen Institute). Travel was generously supported by the Allen Institute. Thank you, Allen Institute! CPCB

[2017] Learning to Explain Non-standard English received an oral presentation at ACL-IJCNLP 2017 in Taipei. UCSB College of Engineering · CS

Selected Talks & Professional Visits

  • SBI2 2024 & 2025, Broad Institute of MIT and Harvard, poster on LivecellAction and short talk on LivecellX (self-funded; attended in person). CPCB
  • CytoData 2023, Allen Institute, poster session (self-funded, with a competitive travel award from the Allen Institute). CPCB
  • MC2 Center for Cancer Biology, invited online talk (1 hour) on LivecellX. CPCB
  • Cancer Biology Program Trainee Progress Report, 30-minute research presentation. CPCB
  • ACL-IJCNLP 2017, Taipei, oral presentation of Learning to Explain Non-standard English. UCSB College of Engineering · CS

A full list of talks, awards, and service is on the Service & Activities page.

Publications

See the full publication list on the Publications page.