About Metaprise
Metaprise is a New York-based AI infrastructure company.
We are building an enterprise-grade Agent Operating System—a foundational infrastructure that enables AI agents to autonomously perform tasks within an enterprise, record every decision, and meet stringent compliance audit requirements.
Our clients are primarily US financial institutions and healthcare systems—in industries where compliance is not optional, and every automated decision must be traceable and auditable.
We are not a chatbot company. Nor are we simply wrappers around GPT applications. We are building next-generation AI infrastructure.
\n\n
¥100,000 - ¥200,000 a month
1、Company technology strategy and co-construction with the founding team
Partner with the CEO and founding team to define company-level technical strategy; own Metaprise's long-term technical defensibility as an AI infrastructure company.
2、Agent Operating System Core Architecture at the Helm
Own the core architecture of the Agent Operating System: agent runtime, state management, multi-agent orchestration, reasoning plane, tool use and function calling frameworks, and failure recovery.
3、Intra-Loop Governance and Auditable Inference System
Design and implement end-to-end traceability for every reasoning step, tool call, context retrieval, retry, and handoff. Build audit capability that allows financial and healthcare customers to replay, explain, and defend every automated decision.
4、Bounded Autonomy and Decision-Making Authority Model
Define autonomy tiers (read / recommend / draft / execute) and design human-in-the-loop approval gates and escalation paths. Ensure high-risk actions stay within bounded, governable scopes.
5、Bounded Autonomy and Decision-Making Authority Model
Treat every agent as a non-human identity with its own access policy: least-privilege permissions, short-lived credentials, context-aware authentication, tool-level access control, and environment isolation.
6、Enterprise-grade reliability and regulated industry compliance
Establish SOC2 Type II, HIPAA, GDPR and financial-industry compliance frameworks. Build high-availability, strongly-consistent, fully-observable, disaster-recoverable enterprise-grade SRE practices.
7、LLM Multi-Model Orchestration and Inference Cost Governance
Design the LLM gateway (routing, fallback, rate limiting, caching), dynamically select among OpenAI, Anthropic, Qwen, or self-hosted models per use case; systematically govern inference cost, latency, and quality.
8、Evaluation System and Data Flywheel
Build evaluation systems for agent performance and correctness, regression testing, and production shadow evaluation; mine gold samples and edge cases from customer usage to drive continuous improvement.
9、Establishment of Shenzhen R&D Center and Technology Culture
As the top R&D leader in Shenzhen, build the engineering organization from the ground up: hiring, org design, talent development, and engineering culture.
10、Sino-US Dual Headquarters Collaboration and Strategic Client Connection
Align closely with the New York HQ on product, sales, and customer success; represent Metaprise in technical due diligence and architecture conversations with CIO/CISO/Head of AI-level executives at strategic customers.
\n
We want to be honest with you about what joining Metaprise means at this stage—because this isn't a role for everyone.
Creative Leader/Technical Partner Level: You're not filling out a technical management position; you're defining it. From technical architecture to organizational structure, it's your canvas.
Directly shaping the technology roadmap for next-generation AI infrastructure: Agent OS is an undefined category, and your architectural decisions will influence its technological trajectory for years to come.
Real clients, real scenarios: Leading US financial institutions and healthcare systems are our clients, and their needs for auditability, compliance, and controllability are real and urgent—meaning any architectural decisions you make can be quickly validated.
Core Partner Level Rewards: A highly competitive compensation package plus equity incentives for the core team. The value you create will be reflected in the company's value.
Dual Headquarters Collaboration in China and the US: Working closely with the New York headquarters, you'll have extensive exposure to the US startup ecosystem and cutting-edge AI practices.