About
Researcher in formulation.
Engineer in practice.
I am Zinuo (Henry) You, a Computer Science PhD candidate at the University of Bristol. I lead modelling research and implementation across UKRI-funded projects spanning generative video, efficient inference, continuous-time models, and evidence-constrained AI systems.
How I work
Across the whole
research system.
My projects typically begin with a modelling question—how a causal video model drifts, where a low-step ODE sampler should spend its evaluations, or how irregular time gaps should enter a latent trajectory. I then build the evaluation and software contracts needed to test the idea without blurring method, data, and protocol.
That means working from derivation and experimental design through PyTorch implementation, multi-GPU training, Slurm/HPC workflows, programmatic evaluation, and release-quality Python packages. I care about negative results, limitations, and strict failure modes because they make research easier to trust and extend.
Education
PhD, Computer Science
University of Bristol
MSc, Electronic & Electrical Engineering
University of Sheffield · Distinction
Bachelor’s, Electronic Science & Technology
Southwest University
Industry
Automotive Engineer Intern
Qualcomm · Real-time perception, INT8/NPU deployment, operator compatibility, and performance diagnosis.
Technical strengths
- Transformers, DiT/MMDiT, causal video
- Diffusion, flow matching, ODE/SDE solvers
- LoRA/PEFT, SFT/DPO, vLLM
- PyTorch, H100, Slurm/HPC
- Evaluation, reproducibility, release engineering
Contact
Open to research scientist
and applied scientist roles.
I am especially interested in teams working on generative video, world models, efficient inference, or research systems where rigorous evaluation matters.