This job is about being a key architect in the 'Picks and Shovels' AI Platform team at Zendesk. You will have the opportunity to build a Digital Nervous System that allows probabilistic LLMs to function as reliable, deterministic employees. The team values collaboration and innovation, working together to solve complex challenges in the enterprise software landscape.
You'll be responsible for
🔧
Building durable agent runtime
You will leverage agentic AI frameworks to build 'ReAct' type loops that are resilient to failure, solving problems around long-running workflows, distributed state management, and async task execution.🔗
Driving integration fabric adoption
You will help drive our adoption of the Model Context Protocol (MCP), building the 'MCP Servers' and 'Clients' that allow our agents to dynamically discover tools and negotiate capabilities with external systems.📚
Engineering context and memory
You will build the 'Memory Manager' for our agents, optimizing RAG pipelines, vector storage, and eviction policies to maximize reasoning fidelity while minimizing token costs.