Founding Engineer

Hobbes•San Francisco, CA

2d•Onsite

About The Position

Hobbes is a self-improving agent that runs product demos today and full sales conversations soon. The team today consists of the CTO and CEO. This role would be the second engineer, focusing on real-time voice, generative UI that redraws itself based on who's watching, custom LLM orchestration, and eval loops that improve performance. The roadmap is not yet defined and this role will contribute to its creation. Hobbes aims to capture the valuable conversations a company has with a buyer, which are often lost at the 'request a demo' stage on B2B websites. Phase 1 involves running real, conversational product demos 24/7. The long-term vision is a self-improving agentic layer that manages the revenue side of a business, with every prospect conversation logged, measured, and fed back into the product for continuous improvement. For example, if five prospects indicate a need for Hubspot integration, the agent would autonomously create the necessary PR and update the prospects. The company has secured $6M in funding and has tripled in size since February, primarily through inbound leads. Early customer successes include one company doubling revenue in 8 weeks and another closing a deal within 2 hours of deployment.

Requirements

Shipped LLM systems to production.
Ability to discuss production issues and solutions.
Ability to move fast without creating hidden problems.
Experience killing features that were not working.
Takes ownership without being asked.

Responsibilities

Build the agent, including LLM orchestration, conversation flow, tool use, and voice capabilities.
Develop generative UI that adapts in real-time to prospect interests.
Implement the eval loop for automated demo improvements.
Create the platform for customers to configure, review, and tune their agent.
Handle all aspects of early-stage development, including CSS, migration scripts, and customer debugging.