Careers at Rifa AI

Building an AI agent is easy. Earning an enterprise's trust in it is not.

We turn written procedures into voice agents that act inside a client's systems and stay inside the limits they set. Every release is evaluated. Every call holds up to a regulator.

ShippingLive on enterprise collections queues
Remote-firstIndia and US, async by default
Seed-stageBacked by Seaborne Capital and operator angels
How we work

A small team with unusually large surface area

We are a handful of engineers serving enterprise clients in production. That ratio is deliberate. It means the person who designs a capability is the person who watches it run.

Own the problem, not the ticket

Engineers here talk to clients, watch their agents run on live queues, and carry a problem from the first call to the metric moving. Nobody hands you a spec.

Evaluation is the craft

Model capability is not our bottleneck; knowing whether a change made an agent better is. If you enjoy building the ruler as much as the thing being measured, you will like it here.

Regulated by design

Our agents work where a call can be read back by a regulator. Constraints are not friction to route around. They are the product, and getting them right is the interesting part.

Ship to production, fast

Small team, paying enterprise clients, real volume. Work you merge this month is on live conversations this quarter, and you will see exactly what it did.

Benefits

What you actually get

Remote across India and the US

Async by default, with deep-work hours protected and overlap kept deliberate rather than accidental.

Health cover for you and family

Comprehensive insurance, plus the flexibility to handle life without filing for permission.

Learning budget

An annual budget for courses and conferences, plus the time to actually use it.

Equipment and setup allowance

The machine you want, the desk you need, and the tooling budget to stop fighting your environment.