Founding Engineer - Agentic

Lack of connectivity is the most fundamental problem in the physical economy.

The physical economy only works as a network of interconnected companies, but it still runs on systems that were never designed to talk to each other. Every time two companies need to exchange data, someone builds a custom integration by hand. Business analysts, integration engineers, QA. All manual. Every time. Google EDI if you haven't. It's wild.

A lot of companies have tried to fix this. Nobody has. We think AI agents are the way to finally connect the physical economy.

Temper is an agentic integration platform for the real economy. Specialized AI agents handle the full lifecycle: gathering requirements in plain English, generating integration code, testing it, deploying it, and keeping it alive in production. When something breaks, the agents fix it.

We have paying enterprise customers. Real integrations running in production right now. There's more demand than we can keep up with.

This is the part most companies are still hand-waving about. You'd actually build it.

Our agents do real work with real consequences in production systems. Your job is to own the systems that make those agents good:

• A context and self-learning system. Agents that get smarter with every integration. A growing knowledge base of undocumented systems that compounds over time, so the platform connecting company #1,000 is far better than the one that connected company #1.
• Self-healing workflows. Agents that detect failures in production, diagnose them, and recover autonomously. No human in the loop for the things that shouldn't need one.
• A meta-harness. The harness our agents run on, plus the meta-layer above it that lets us build, evaluate, and evolve new agents fast. The thing that turns “we need an agent that does X” into a shipped, measured agent in days.

Nobody has shipped this kind of system at production scale yet. You'd be one of the first.

You'll be our founding agentic engineer, working directly with Nick (CTO).

• Own the agent layer end-to-end: context, self-learning, evals, the harness, the meta-harness.
• Set the patterns for how every agent at Temper gets built, measured, and improved.
• Build the team. You'll help us hire it, pulling strong builders from your network, and grow into leading the people who own agentic systems.
• Work as a peer to the founders. We need your opinion, not just your code.
• Move fast and experiment relentlessly. Most of what works here hasn't been invented yet. Your job is to find it.

The agent-building paradigm gets rewritten every few months. Harness-centric building, context engineering, eval-driven development: the playbook is being written right now, in public, and we're writing parts of it.

If you want to start your own company in agentic software, this is the launchpad. You'll work on the most novel agent tech that exists, in production, with real consequences, and build the instincts the next wave of agent companies will be built on.

• Daniel

  Grew up in his family's trucking business. Built the freight product at Europe's leading eCom logistics startup, Hive, from 0 to $M and taught himself to code well enough to ship a real product in under a year.

• Heinrich

  Ran 100+ person warehouse operations at Amazon, then consulted on system integrations at KPMG. Built Hive's largest revenue line from zero.

• Nick

  A hacker at heart: built Lilith at 16, led an offensive security program for a nation-state, entrepreneur for 8 years. In eCom, he led the tech that scaled a company from 0 to $100M in 3 years.

Together, Daniel and Heinrich built Europe's leading eCom logistics startup from pre-seed to 400+ people.

We care about the craft. We move fast and want to build something we're proud of.

• You've built agentic software in the last year. This field moves monthly, so we need someone who's been in it recently, not someone who read about it.
• You live in the newest paradigm of harness-centric agent building. Context engineering, harness design, eval-driven iteration: this is already how you think.
• You understand agent tracing and evals deeply. An agent you can't measure is an agent you can't improve.
• Implementation engineering, not research. You ship working systems. We're not hiring an ML or data-science background. We're hiring a builder who makes agents work in production.
• You live on X (or wherever the frontier actually gets discussed) and you experiment constantly: new models, new tools, new techniques, the day they drop.
• A high desire to work on the most novel tech that exists. If it's been done before, you're less interested.
• You code with AI agents and lean into it, but have the taste not to ship slop.
• Strong technical opinions, held loosely. You change your mind when the evidence says so.
• You don't wait to be told what to do.
• You want a true early-stage, in-person experience in SF: fast-paced, high intensity.

• Top-quartile pay
• Founding equity
• A top-tier office in San Francisco
• Equipment for maximum productivity: M5 MacBook Pro & more
• Full health coverage, 401(k) match

1. 1
   Screen20 min

   Quick chat, mutual fit.

2. 2
   Deep dive45 min

   Your career, what you've built, how you think.

3. 3
   Technical75 min

   Design a system with us. No leetcode.

4. 4
   Paid work day

   A real task from our backlog, in person.

5. 5
   Offer

   We move fast. Expect under two weeks from first call to offer.