AI Memory Systems
How to give AI persistent, searchable context across sessions.
See MeMesh →
RESEARCH LAB
AI Application Research Lab
We investigate how AI can solve real-world problems — then build tools that prove it.
A startup team from Taiwan.
Our Mission
Most AI tools are built on assumptions. We start with research: deep dives into real-world workflows, systematic problem analysis, and rigorous testing. Only then do we build — creating production-grade tools that developers can trust.
Team
Kai Hsiang Tseng
Founder · Hsinchu, Taiwan
Serial entrepreneur, 20 years across software, mobile, and industrial systems. CS from University of Massachusetts Boston. Every AI system I built hit the same wall — sessions start from scratch, decisions made last week are invisible to agents running this week. I built MeMesh to fix this for my own workflow. 50,000+ developers adopted it. Now I'm building the cloud sync layer that makes AI memory a team capability — not just a personal one.
Arthur Liu
Technical Advisor & Partner · AI Systems · Ph.D. Candidate, NTU Biomedical Technology
AI systems researcher and engineer with deep expertise in applied machine learning and computational modeling. B.S. in Computer Science and Information Engineering, National Taiwan University. Ph.D. candidate in Biomedical Technology, NTU. Focused on building AI infrastructure that bridges research and production — the same problem PCIRCLE.AI solves at the memory layer.
T.H. Chen
Partner · Machine Learning & AI Models
Machine learning researcher specializing in AI model development and training. B.Eng. in Information Technology Engineering, York University, Canada. Brings deep technical knowledge of model architecture and optimization to PCIRCLE.AI's research-to-production pipeline.
Vincent Liang
Business & Growth Advisor
Serial entrepreneur with proven exits in e-commerce. M.S. in Management Information Systems, University of Calgary, Canada. Investor and operator across live commerce, crypto, and AI-native applications. Specializes in applying emerging technology to established business verticals — bringing the commercial perspective that turns PCIRCLE.AI's research into real-world enterprise adoption.
Research → Build → Validate → Share
1
Identify real-world problems in AI development
2
Build solutions with production constraints in mind
3
Validate with 50,000+ real users before 1.0
4
Open-source everything, publish learnings
Current Research Areas
Three interconnected research threads, each producing tools that work in production.
Workflow Intelligence
AI-powered diagnosis of complex administrative processes.
See the tools →Impact Assessment
Quantifying social and environmental outcomes in Web3 systems.
Explore projects →Partner With Us
We work with organizations to apply our research to domain-specific problems. If you're facing challenges in AI deployment, workflow automation, or impact measurement, let's talk.