You own more of the problem.
The team is small, so the surface area is large. You take a problem from the reality it lives in, through the system that solves it, to the people whose work it changes. Range is how you scale up, faster than a divided team allows.
Work to be done.
You build full-stack: a Python or Go backend, a modern frontend, the LLM orchestration and retrieval that make an agent useful, and the data pipelines that feed it. You own the deployment as well, from containers and CI to the cloud it runs on and the monitoring that watches it. Much of it sits in regulated settings, so reliability and a clean audit trail are designed in, and you carry a problem from the client conversation to the live system.
You design how a system asks for what it needs, earns trust as it goes, and holds a brand's voice through every exchange. The conversation and the decision are the product, so the thing you shape is behaviour. You sit beside the engineers and test against real use, revising on how people behave. It is a young surface, and you help set how it works.
You sit within the client's work and map how it really runs, which is seldom what the org chart says. You shape what gets built, design the rollout so the system is one people still choose in month six, and work with senior stakeholders from the first session through to live. The judgement you build about how organisations actually adopt is rare, and it travels with you.
Because the team is small, the disciplines meet. An engineer sits in the client conversation. A designer reads the evaluation data. Across a year you touch several sectors, from manufacturing to field healthcare, which is its own education. The people who go furthest follow a problem past the edge of their discipline, and there is room to do it.
Range is how you scale up.
A narrow role hands you one slice and keeps the rest at arm's length. Here you carry the whole problem, from the reality it lives in to the people it changes, across sectors and disciplines. That is the kind of work that compounds into judgement, and it compounds fast.
We build agentic AI that does the work and keeps the record of how it gets done, in production and across demanding and regulated sectors. The team stays small on purpose, so the range stays real and the work stays yours.
What the work asks of you.
- Initiative that does not wait to be asked.
- A bias to build, and the nerve to take a risk.
- Integrity that holds when nobody is checking.
- The honesty to review your own work and change course.
- A mind that spots patterns across fields that do not obviously connect.
And the ability to say it plainly.
How we work
We work across Pune and Amsterdam. In-office, hybrid and remote all sit comfortably with us. What matters is the work and the people doing it, wherever they are.
Where do you want to go next?
Tell us who you are and why you'd like to work with us. The rest of the conversation is about the work.
- 1You send it.A few details, and why you'd want to work with us.
- 2We read it ourselves.Sushrut and Nikita, not a filter.
- 3We write back if there's a fit.And it starts with the work.
Your details are handled under our information-security and data-protection commitments, and used only to consider you for work with us.