Christopher Wu
Head of ML Engineering
Former ML tech lead at Meta with seven years of experience.
Our investors
Our seed round was led by Point72 Ventures, with participation from Qualcomm Ventures, Samsung NEXT, and other investors supporting our mission.
$12M seed round
Announced July 2025
Lead investor
Qualcomm Ventures
Samsung NEXT
Also backed by Salience Capital, SCB 10X, Jinqiu Capital, Aiconic Ventures, and Sixty Degree Capital, alongside angel investors including Lip-Bu Tan and Abhay Parasnis.
Our team
Zheqing (Bill) Zhu
Founder and CEO
Bill leads Pokee AI as Founder and CEO, bringing frontier AI to infrastructure enterprises control. He was previously Meta AI’s Senior Staff Research Lead Manager and holds a PhD in reinforcement learning from Stanford University.
Core engineering & research team
Yi Wan
Founding Research Scientist
Former researcher at Meta AI. Conducted PhD research with Rich Sutton at the University of Alberta.
Michael Cai
Head of Product Engineering
Seven years of experience leading product engineering teams across Meta and Agoda.
Ruihao Zhu
Principal Research Scientist
Assistant Professor at Cornell University with a PhD from MIT.
Qiwei He
Senior Research Scientist
Experience at FocusAI, acquired by Qualcomm, with a focus on model training and silicon optimization.
Chang Zen Tee
AI Infrastructure & Inference
Experience optimizing AI inference at TSMC and ByteDance.
Want to build with us? See open roles
How we build
We develop the model and the agent platform together, around the realities of enterprise work.
Bring AI to the data.
Some work needs to stay in a private environment. We design for that from the start, with deployment in a customer’s VPC, on premises, or on device.
Keep the context.
Decisions often depend on a full record. Our long-context architecture helps agents work across extensive documents and histories while retaining the connections between them.
Make the work usable.
We build toward completed work: a report a team can review, a workflow it can run, or a tool it can use. Planning and execution belong together.
Pokee-Isaac
The model
A 28B-parameter agent model with a native 10M-token context, built for reasoning and tool use across large bodies of information.
Evaluate through the API. Deploy in your environment.
PokeeClaw
The agent platform
The platform that turns models into working agents, with planning, memory, tool use, and policy controls connected to your systems.
Your workflows. Your infrastructure. Your control.
News about Pokee AI
All insightsIntroducing Pokee-Isaac 28B
Product launch · Pokee AI on X
Our agentic model brings native 10M-token context to enterprise work, with deployment in your VPC, on premises, or on device.
Agentic AI on Snapdragon X Series PCs
Qualcomm · August 2026
Qualcomm features Pokee among the applications bringing AI agents onto local PCs.
Named to CB Insights’ AI 100
CB Insights · May 2026
Pokee joins the 10th annual list, with recognition in enterprise AI.
Pokee AI raises a $12M seed round
GeekWire · July 2025
Coverage of our funding led by Point72 Ventures and our work on AI agents.
Let’s put AI to work
in your world.
Meet the team. Explore what’s possible for your organization.