Applied AI Engineer

Orpex; San Francisco (CA); On-site; USD 150,000 - 250,000; Full time

Benefits offered by this job

Relocation and visa sponsorship

In-person in San Francisco

Job summary

Orpex seeks an Applied AI Engineer to build the agentic infrastructure powering real company operations in San Francisco. You will develop memory and contextual systems, design evals, and ship robust agents that perform tasks reliably for customers.

You will own the loop: build it, measure it, break it, fix it, and teach the system to learn. Expect hands-on infra work, orchestration, customer interaction, and helping grow the engineering team.

Qualifications

• Have shipped LLM agents that ran unattended for real users.
• Built eval harnesses used to decide what ships.
• Memory and retrieval that work on messy operational data, not demo data.

Responsibilities

• Ship a working agent system from day one.
• Own the loop: build, measure, break, fix, and improve learning.
• Handle infra/orchestration, customer calls, and hiring the next engineer.

Skills

Agents in production

Evals you trusted

Memory and context

Intensity

Job description

ABOUT ORPEX

The next generation of service companies won't scale through headcount. They'll be built by tiny teams orchestrating AI workforces with 1000x the output. Orpex is building the infrastructure they run on: agentic workforce, company brain, human trust layer, one cockpit to run the business.

We just raised an oversubscribed 2.5M angel round in 7 days and are moving at rapid speed.

APPLIED AI ENGINEER

Everyone is building agents. Almost nobody is building agents a real company can run on.

You own the agentic infrastructure of Orpex: the workforce that executes, the company brain that remembers, and the evals that prove it works. Agents that run the same task for the fiftieth time as reliably as the first. Memory that holds months of client context. Approvals that pull a human in at exactly the right moment. You decide how agents plan, remember, fail and recover. You'll ship to production in your first week.

WHO WE'RE LOOKING FOR

• Agents in production. You've shipped LLM agents that ran unattended for real users. You know the model is the easy part.
• Evals you trusted. You've built eval harnesses that people actually used to decide what ships.
• Memory and context. Retrieval that works on messy operational data, not demo RAG.
• Intensity. You force things into existence. Bonus if you've built agent tooling in the open or founded something.

YOUR FIRST MONTH

• Ship from day one. An agent running for a real company in week one.
• Own the loop. Build it, measure it, break it, fix it, make it learn.
• Do what needs doing. Infra, orchestration, customer calls, hiring the next engineer.

WE OFFER

• $150–250K USD base + meaningful early equity
• San Francisco, in person. Relocation and visa sponsorship covered.
• Pace and purpose. Small team, high standards, high trust, no politics.

PROCESS

1. 30-min intro call
2. 90-min deep dive on an agent system you shipped: what broke, what you measured
3. Technical task
4. Offer the same week

If you want to build something truly generational and are ready to make it your life's mission, this is for you!