Lead Engineer, AI Agent Systems

About the role

Patsnap is hiring a technical leader to build and refine the infrastructure behind our AI agent systems. You will oversee the development of complex frameworks that manage reasoning, context, and execution for knowledge-heavy tasks. This role requires a balance of hands-on coding and high-level architectural design to ensure our systems remain reliable and scalable.

Key facts

What you'll do

• Architect the infrastructure for agent systems, focusing on the execution engine, reasoning orchestration, and core capabilities.
• Build the runtime for agent loops, including task decomposition, concurrency management, and failure recovery.
• Develop systems for context management, such as token-budget governance, information compression, and evidence tracking.
• Create secure foundations for agent operations, including sandboxing with Docker and Kubernetes, memory stores, and MCP hubs.
• Write production code for critical modules and guide the team through technical reviews and architectural decisions.
• Implement observability and diagnostic tools to monitor agent performance and reliability.

Requirements

• Minimum of five years of professional software engineering experience.
• Previous experience as a Tech Lead, Staff Engineer, or equivalent role with a team of at least three people.
• Proven track record of building complex systems that go beyond basic API integrations or CRUD applications.
• Advanced Python programming skills.
• Expertise in distributed systems, asynchronous execution, and streaming responses.
• Deep hands-on experience in at least two of the following: Agent Execution Engines (reasoning loops, planning), Context Orchestration (token management, input shaping), or Agent Foundations (sandboxing, memory, security).

Nice to have

• Experience with multi-model routing and provider integration.
• Background in building automated evaluation systems for AI outputs.

Skills & tools

• Python
• Docker
• Kubernetes
• AST-based controls
• MCP (Model Context Protocol)
• Distributed systems architecture

Practical notes

This position is based in Shanghai. Candidates must demonstrate the ability to manage the trade-offs between system latency, cost, and output accuracy in a production environment.