Founding Engineer

Hobbes; San Francisco (CA); On-site; USD 180,000 - 230,000; Full time

Benefits offered by this job

Equity: Founding-caliber

Job summary

Hobbes, located in San Francisco, is seeking a second engineer to help build an intelligent agent that runs real-time voice and generative UI. You'll contribute to LLM orchestration and the eval loop, making demos smarter without manual intervention. With $6M in funding, Hobbes focuses on enhancing customer interactions through self-improving agents.

The position requires a proactive individual with experience in deploying LLM systems to production.

Qualifications

• Experience shipping LLM systems to production and handling issues.
• Ability to work quickly while avoiding hidden problems.
• Demonstrated ownership of tasks without being prompted.

Responsibilities

• Build an intelligent agent with LLM orchestration and conversation flow.
• Develop a generative UI that adapts in real-time.
• Ensure the eval loop makes demos smarter automatically.

Skills

LLM orchestration

Conversation flow design

Real-time UI development

Debugging

Job description

Hobbes is a self-improving agent that runs product demos today and full sales conversations soon. Arjun (CTO) and Chirag (CEO) are the team today. You'd be the second engineer: real-time voice, generative UI that redraws itself based on who's watching, custom LLM orchestration, eval loops that actually move the needle. The roadmap isn't written. You'll help write it.

About Hobbes

A company's most valuable asset is the conversation it has with a buyer. Most B2B websites lose that conversation at the "request a demo" form. Hobbes runs real, conversational product demos 24/7. That's Phase 1. The real bet is a self-improving agentic layer that runs the revenue side of a business.

Every conversation a prospect has with the agent is logged, measured, and fed back into the product. Five prospects tell Hobbes they want to buy but need Hubspot integrated as a non-starter? Hobbes creates the PR autonomously and updates those prospects that it's ready to go-live.

We have secured $6M in funding and have tripled since February, mostly on inbound. One customer 2x'd their revenue within 8 weeks of deploying Hobbes. Another closed a deal within 2-hours of Hobbes going live on their site.

What you'll build

• The agent. LLM orchestration, conversation flow, tool use, voice.

• Generative UI that adapts in real time to what the prospect cares about.

• The eval loop. Demos get smarter without manual intervention.

• The platform customers use to configure, review, and tune their agent.

• And yes: CSS, migration scripts, customer debugging. Early-stage means all of it.

What we're looking for

• You’ve shipped LLM systems to production and can talk about what broke and what you did about it.

• You move fast without creating hidden problems. You've also killed features that weren't working.

• You take ownership without being asked.

Comp

• Base: $180-230K

• Equity: Founding-caliber

• Location: In-person, San Francisco

Process

4 rounds: Arjun, Chirag, async case study, paid 3-day work trial in the office. About 2 weeks from first conversation to offer.