All open roles

Data Scientist (LLMs / Conversational AI)

Location: Tel Aviv

Role Type: Full-time, hybrid

Our Stack: Python, LLMs & fine-tuning, RAG, evaluation pipelines, messy conversation & business data at scale

Overview

We build the AI-native layer for enterprise customer service. We ingest a company's service data and learn how their business actually runs. We turn this data into AI agents that answer the phone, assist human agents in real time, and surface the insights that show what's really happening in the business.

The Technical Challenge

  • The Grounding Problem: Sits at the heart of the job. Taking a model that's never seen a business and grounding it in that business well enough to answer real customers correctly, safely, and in its own voice.
  • The Signal Problem: Brings its own difficulty. Turning a thousand messy conversations into evaluation sets, insights, and something a human can actually act on, at scale, and without hand-labeling everything.

The Role

You own problems end to end. From the ingestion and grounding pipeline to the evaluation harness to the insights the customer sees. Small team, short path from idea to production, nothing between you and what ships.

You'd Fit If You:

  • Are an experienced data scientist, comfortable with LLMs, RAG, and evaluation.
  • Take ownership: You carry a problem from idea to production and stand behind it.
  • Like to tinker: You like to play with messy conversation and business data.
  • Work independently: You figure out what needs doing without being handed it.
  • Learn and ship fast: Learn fast and ship fast.

Let's Talk

If building AI that talks to people, supports the humans behind them, and understands the whole business sounds like your kind of challenge, let's talk.

Apply for this role