About the role
TextYess is an Italian-born AI-first CRM scale-up building autonomous AI agents that sell, support, and retain customers for B2C brands. The company has raised over 3 million euros and works with names like Ducati, Pittarosso, MC2 Saint Barth, and Doppelganger. This AI Engineer role sits on the AI team and focuses on the agentic runtime for proactive agents, covering context building, tool and retrieval orchestration, scoring, and guardrails. You will own production components from day one.
The work is deeply hands-on and evaluation-driven. You will tackle grounding failures where agents misreport stock, combine exact-match and semantic retrieval for SKUs and variants, and maintain multilingual quality across Italian, English, and other languages. Latency and cost profiling, prompt and model experimentation, and improving a golden-dataset harness tied to Langfuse traces are all core responsibilities. You will.
Highlights
- Own production AI components from day one on a lean, fast-moving team
- Work across RAG, retrieval orchestration, evals, and agent guardrails
- Use Langfuse traces and a golden-dataset harness to predict production quality
- Collaborate with international talent on real B2C brand deployments
This recap is dataskew's editorial summary, not the company's copy.