What you'll do

• Design retrieval and orchestration pipelines over products, orders, suppliers, and freight data.
• Build evaluation harnesses that keep assistant answers precise as the product evolves.
• Ship assistant-driven workflows into production modules with the product engineering team.
• Tune latency, cost, and quality trade-offs across model providers.
• Prototype fast, measure honestly, and kill what does not work.

What we look for

• Three or more years of production software engineering, with recent LLM application work.
• Hands-on experience with retrieval pipelines, structured outputs, and evaluation tooling.
• Strong Python or TypeScript engineering fundamentals: this is a software role first.
• Judgment about when AI is the right tool and when a form field is.
• Bias for shipping measurable improvements over polishing demos.

Nice to have

• Experience with agentic workflows acting on ERP-like systems.
• Published evals, open-source tools, or writing about applied LLM work.
• Familiarity with supply chain or commerce data.

Benefits

• Competitive salary and meaningful equity
• Health, dental, and vision coverage
• 25 days of paid time off
• Annual learning and conference budget
• Home office setup stipend
• Quarterly team weeks in São Paulo