Open role
AI Builder
If you've ever designed an agent's failover by thinking “what does this system do when the LLM returns the wrong tool call three times in a row” — we should talk.
Most AI projects fail at deployment, not at modeling. The bottleneck is engineering judgment — knowing what to abstract, what to make autonomous, what to keep human-in-the-loop, what to monitor. We build agents that actually run for real operations companies. We're looking for someone who designs AI systems exceptionally well.
What we look for
- 01You've shipped a multi-step agent system to real production. It failed in interesting ways. You learned from the failures.
- 02You write the prompt, the eval, the failover and the dashboard yourself. No hand-offs.
- 03You have informed opinions on agent design — framework vs. raw, model choice per task, when to split a problem into more agents vs. fewer.
- 04Most of all, you know when to use AI and when not to.
Apply only if you're okay with all of this
- You can work in person from our HSR Layout office.
- This is the kind of work where you'll think about a system over the weekend because you want to, not because anyone asked. If you're optimizing for clean work-life boundaries, this isn't the right fit — but you won't be working alone.
- This is not a typical engineering or FDE role. It's for someone who already designs AI systems exceptionally well. Strong backend engineering is the floor, not the bar.
What you'll do
Bala leads the business conversations and acts as the FDE for customers. He'll teach you the business and how to break problems down from a business point of view. You bring the deep technical skill to build reliable, scalable agentic systems.
- Work directly with Bala Panneerselvam — visiting customers together, diagnosing problems together, designing solutions together.
- Design the agents and systems that customers use in production.
- Build the platform and infrastructure that runs the studio internally — about 40 percent of our time goes into internal tools and mechanisms today.
- Work with AI models every day, for coding and non-coding work alike.
How the interview works
- 1
A phone call
We understand what you're looking for, and what we're looking for.
- 2
First in-person conversation
Your background, your mental models, the way you think through problems.
- 3
A large assignment
A few days of real work. You're welcome to use every AI tool you have to arrive at the solution.
- 4
Final in-person conversation
We close the loop together. Expect the whole process to take about a week.
Why this role
We'll have a compensation conversation during the interview. Beyond that, if you're optimizing for a long-term career and willing to put in the effort:
- Put in the effort, and by 2028 you'll be one of the foremost experts in deploying AI inside large-scale enterprises.
- You become a builder — not just an engineer with skills, but someone who understands a business, breaks it down to process, and applies the right technology.
- A front-row seat to some of the fastest-growing businesses in the world.