Founder & CEO · Lawless, Inc. Currently building

Zhiyuan
Ren任志远

Zhiyuan Ren
Portrait · MMXXVI

Building Legal AI from a research background in computer vision, multimodal models, and agents.

Founder & CEO of Lawless — a vertical AI company for the legal stack. Previously a research scientist at Megvii, SenseTime, Baidu, OPPO, and Tencent. CVPR / NeurIPS author with 500+ Google Scholar citations.

Now
Lawless /01
Citations
500+
Top venues
CVPR · NeurIPS
Based in
San Francisco
Scroll
01 Identity

A researcher-turned-founder. Trained in computer vision, multimodal learning, and language agents — and now full-time on the hardest, least-glamorous corner of applied AI: law.

The bet is simple. The next decade of legal practice will be rebuilt on top of trustworthy delegation. Lawless is how I plan to build the substrate.

Now
Founder & CEO, Lawless, Inc.
Patent risk · Lawyer-assistant agents · TRO-class workflows
PhD
Michigan State University · AI (left to build Lawless)
Advisor: Prof. Xiaoming Liu
B.Eng.
Huazhong Univ. of Science & Tech.
Qiming Honor College, CS · Exchange year at UCSB
Prev.
Research Scientist · Megvii · SenseTime · Baidu · OPPO · Tencent
02 Lawless · What I'm building

Beyond demo-ware.
A trustworthy legal stack.

Lawless is a vertical AI company for the legal industry. We pick problems where automation only ships if it earns a lawyer's trust on the first try — patent risk, delegation-grade assistants, and frontier TRO-class workflows. Three product pillars, one principle: accuracy before flash.

P1 / Core
Patent Risk Analysis
In production · with law firm partner
End-to-end engine for patent infringement and freedom-to-operate analysis. Iterating directly with Attorney Zhang Chao's firm; integrating with the Zhihuiya patent database for full-corpus recall.
Live · iterating
P2 / Flagship
Lawyer-Assistant for PWC
Delegation-grade · beyond Harvey
A delegation tool that out-automates Harvey while putting trust first. The bar isn't "draft something plausible" — it's "complete the task to a senior lawyer's standard, with auditable reasoning".
Active build
P3 / Frontier
TRO-class Workflows
Exploration · 0→1 motions
Consumer-side TRO settlements, structured consultation outputs, and AI-native company management. Frontier surface area where the legal interface and the agent interface collapse into one.
Exploring
03 Selected research

Vision, multimodal,
and agents.

CVPR
2024
TIGER: Time-Varying Denoising Model for 3D Point Cloud Generation with Diffusion
Zhiyuan Ren, Minchul Kim, Feng Liu, Xiaoming Liu
PDF Code
NeurIPS
2023
ChatGPT-Powered Hierarchical Comparisons for Image Classification
Zhiyuan Ren, Yiyang Su, Xiaoming Liu
PDF Code Slides
CVPR
2024
Distilling CLIP with Dual Guidance for Learning Discriminative Human Body Shape Representation
Feng Liu, Minchul Kim, Zhiyuan Ren, Xiaoming Liu
PDF
CVPR
2023
Hierarchical Fine-Grained Image Forgery Detection and Localization
Xiao Guo, Xiaohong Liu, Zhiyuan Ren, Steven Grosz, Iacopo Masi, Xiaoming Liu
PDF Code
ICASSP
2022
Diffusion Motion: Generate Text-Guided 3D Human Motion by Diffusion Model
Zhiyuan Ren, Zhihong Pan, Xin Zhou, Le Kang
PDF
Full publication list → Google Scholar  ·  GitHub
04 Trajectory

Academia × industry,
into founding.

Education

— 2024 · Left to build Lawless
PhD, Artificial Intelligence
Michigan State UniversityUSA
2019 — 2020
Visiting Scholar, Computer Science
UC Santa BarbaraExchange
2016 — 2020
B.Eng., Computer Science
Huazhong Univ. of Science & TechnologyQiming Honor

Industry · Research Scientist

Five Research Scientist roles across the leading AI labs in China and the U.S. — delivering models, papers, and deployed systems for vision, multimodal, and agentic stacks.

Megvii
旷视
SenseTime
商汤
Baidu
百度
OPPO
U.S. Lab
Tencent
腾讯
+ MORE
05 Contact
For collaborations, partnerships, or just to say hi —
Reach me at
zhiyuanren2200@gmail.com.
Google Scholar ↗ GitHub ↗ Twitter ↗ CV (PDF) ↗