Multimodal patient models
Founder-minded work on systems that combine histopathology, transcriptomics, EHR context, and model infrastructure into something clinically legible.
Computational biology, AI, oncology, writing
Building serious systems for biology, medicine, and the people who have to use them.
Hanson Wen is a Molecular and Cell Biology + Computer Science student at UC Berkeley working across computational biology, AI systems, oncology, and writing.
The work is intentionally split across research, product judgment, and public thinking: multimodal models for biology, stronger evaluation for agent systems, and essays that make the underlying ambition legible.
- Incoming Computational Biology Intern at Regeneron
- Student Researcher at UC Berkeley Sky Computing Lab
- Student Researcher at Berkeley RDI
- Writer behind The Quixotic Method
Three active threads
One site, not one giant homepage.
The landing page stays focused. Deeper detail lives where it belongs: dedicated pages for projects, writing, background, and a resume-backed view of the work.
Scientific tooling that can actually reason
Research spanning agent benchmarks, swarms for biology, and the interfaces needed to turn capability into reliable action.
Public thinking as part of the product
The writing is not decorative. It is where ambition, philosophy, technical taste, and long-horizon judgment get made explicit.
Selected projects
Projects with enough edge to be worth their own pages.
Enso Atlas
A fully local pathology AI system for treatment-response prediction, evidence retrieval, and reporting on whole-slide images.
Top 10 in Google MedGemma Impact Challenge
TrialScope AI
An agentic workflow for extracting decision-useful insight from clinical trial protocols and trial documents.
CalHacks 12.0 Regeneron Track Winner
Personal Genome Analysis
Processed 131GB of whole-genome sequencing data, identified 4.99 million variants, and built a full pipeline for ancestry and pharmacogenomics analysis.
Full-stack genome pipeline
AgentHLE
Humanity's Last Exam for computer-use agents, spanning 63 high-GDP domains and a much tougher standard for useful agent performance.
Benchmark design at Berkeley RDI
Writing
Writing is first-class, not a sidecar.
The site treats writing as core infrastructure. It is how technical judgment, taste, and ambition become inspectable.
Between the windmill and the brick.
Essays on AI, ambition, philosophy, computational biology, and the discipline of choosing problems that can matter for a long time.
For people who want compact signal.
Current experience, education, projects, and awards rendered directly from the resume source.
- Regeneron Pharmaceuticals — Incoming Computational Biology Intern
- UC Berkeley Sky Computing Lab — Student Researcher
- Berkeley RDI — Student Researcher
- Stealth Biotech — Co-founder