About
Minkyu Jeon is a Ph.D. candidate in the Department of Computer Science at Princeton University, advised by Prof. Ellen D. Zhong. His research develops generative and representation learning methods for scientific inverse problems, with a current focus on heterogeneous cryo-EM reconstruction and AI-driven de novo protein design. He is a former summer research intern and contractor at Prescient Design (Genentech), and a former Associate Computational Biologist and visiting graduate student at the Broad Institute of MIT and Harvard.
Research Interests
My research focuses on generative models, representation learning, and 3D reconstruction for scientific inverse problems, with an emphasis on cryo-EM heterogeneous 3D reconstruction.
Another key area of my research is AI for Science. I am currently working on AI agents for scientific discovery, adaptive masked diffusion models, and scalable generative models for vision and proteins. A major current direction is agentic inference-time scaling for de novo protein design, where LLM-guided search, structural validation, and iterative refinement are combined into executable design workflows.
Selected Projects
- Agentic Inference-Time Scaling for De Novo Protein Design: AI-agent workflows for protein binder and scaffold design, with live tracking of Qwen-guided config search, validation runs, and hard-target diagnostics. Project hub / Historical archive
- CryoBench and Heterogeneous Cryo-EM Reconstruction: datasets and amortized reconstruction methods for challenging conformational heterogeneity problems in cryo-EM. CryoBench
News
| Date | Update |
|---|---|
05-2026 | Added the current project hub for agentic inference-time scaling for de novo protein design. |
04-2026 | Preprint released: ConforNets — Latents-Based Conformational Control in OpenFold3. |
02-2026 | Our work, a transformer-based hypernetwork for cryo-EM reconstruction, has been accepted to CVPR 2026. |
12-2025 | Preprint released: CryoHype — Reconstructing a thousand cryo-EM structures with transformer-based hypernetworks. |
12-2025 | Preprint released: CHIMERA — Adaptive Cache Injection and Semantic Anchor Prompting for Zero-shot Image Morphing. |
11-2025 | Started part-time research internship at Prescient Design (Genentech). |
10-2025 | Two papers accepted at NeurIPS 2025 Machine Learning for Structural Biology (1 poster and 1 Oral). |
06-2025 | Excited to begin summer internship at Genentech’s Prescient Design. |
06-2025 | Preprint released: R3eVision — Survey on Robust Rendering, Restoration, and Enhancement for 3D Low-Level Vision. |
05-2025 | Passed the general exam. Now I’m a PhD candidate! |
09-2024 | Our CryoBench paper on cryo-EM heterogeneous reconstruction benchmarks accepted to NeurIPS 2024 (Spotlight). |
08-2024 | Presented CryoBench paper at Flatiron Institute. |
07-2024 | Instructed high school students at AI4ALL Princeton. |
02-2024 | Gave an invited lecture on k-SALSA at CMU MetaMobility Lab. |
09-2023 | Starting Ph.D. journey at Princeton! |
01-2023 | Starting an Associate Computational Biologist role at the Broad Institute of MIT and Harvard. |
11-2022 | Paper on self-supervised representation learning for localization accepted to Information Sciences Journal 2023. |
09-2022 | Paper on saliency-guided point cloud data mixup accepted to NeurIPS 2022. |
07-2022 | Paper on GAN-based privacy preservation of retina images accepted to ECCV 2022. |
09-2021 | Starting as a visiting graduate student at the Broad Institute of MIT and Harvard. |
Teaching
- Instructor, AI4ALL — Princeton University, Summer 2024
- Teaching Assistant, Mathematics for Numerical Computing and Machine Learning (COS302) — Princeton University, Fall 2024
- Teaching Assistant, Artificial Intelligence (COSE361) — Korea University, Spring 2021
- Teaching Assistant, AI Security (AAA712) — Korea University, Fall 2020
- Tutorial Teaching Fellow, Deep Learning — Korea University, Fall 2020
