Portrait of Yangyi Liu

Yangyi Liu

I founded AmadeusBio.ai, where I build AI agents that accelerate biological research.

I trained as a molecular and cell biologist at UC San Diego (B.S. 2025, cum laude) and spent two years at the bench, in yeast synthetic biology and C. elegans neurogenetics, before building AI agents for research. CV

I’m open to opportunities in AI for biology, in research or engineering. Email is the best way to reach me.

Selected work

  1. Oct 2026

    Molecular Animation

    AmadeusBio.ai · open source, Apache-2.0Released v0.2.0

    An open framework that turns a scientific brief into a molecular film with Blender. Every film comes with a production record that states, beat by beat, what is experimental and what is illustrative. The channelrhodopsin film took about two hours from idea to final render.

    The channelrhodopsin film: the light-gated channel behind the 2026 Nobel Prize in Physiology or Medicine, built from PDB 9GO1 and 9GO2, with each beat labelled on screen. It was released two days after the prize was announced. The membrane is procedural, not simulated.
    Describe the film

    A 21-second silent film of channelrhodopsin C1C2, built from PDB structures 9GO1 and 9GO2. It opens on the protein dimer in a lipid membrane as photons arrive, dives into the retinal pocket, and shows retinal twisting from all-trans to 13-cis, an illustrative morph between the two crystal states. Lys132 then swings toward Glu129 to form a salt bridge. A section through the cation pathway first shows it interrupted, as in the crystal, then open in an illustrative conducting-state model as sodium ions cross the membrane. Labels on screen mark each beat as experimental, illustrative, or symbolic.

  2. 2025

    Neuronal morphology genetics in C. elegans

    Jin Lab, School of Biological Sciences, UC San DiegoPoster · 2025

    Undergraduate research on genes that maintain the shape of touch-receptor neurons, using genetic crosses, PCR and Sanger genotyping, and fluorescence microscopy. First-author poster at the 2025 International Worm Meeting.

All work, with roles, limits, and evidence

How I work with AI agents

“When the data disagrees with the story, show what the data says first. Then add the illustrative step and label it.” From the Molecular Animation principles

Measure the science before building on it.
Molecular Animation fetches the PDB entries and measures every claim a film depends on, in code, before anything is animated. Workflow
Agents build; checks decide.
Every build reports its numerical checks. For the channelrhodopsin film, they cover the retinal dihedral path, a salt-bridge distance, the worst bond deviation in the morph, the pathway profile, and ion clearance.
Label what is known.
Each beat of a film is tagged experimental, illustrative, or symbolic, so what comes from a deposited structure stays separate from what is drawn. Principles
Inspect before release.
Before the final render, the stills, a preview contact sheet, and motion statistics that flag jumps and seams are reviewed. A check command re-renders a shot and compares it with the delivered poster.

Outside work

A digital painting by Yangyi Liu: a girl with long brown hair sits hugging her knees in a sunlit corridor.
A digital painting of mine.

I paint and draw, mostly anime-style illustration; a few pieces are on Pixiv. I also read visual novels and feed the crows on campus. More about me