tldr;
At Gentoo, we’re a small team building the agentic commerce platform.
We’re still early, which means every engineer here has a chance to shape not just our product, but redesign the future of e-commerce in the AI-native era.
We are looking for ML engineers who turn open ended shopper questions into reliable systems. You will design retrieval and ranking that surface the right products, craft prompts and policies that keep conversations grounded, and build the data pipelines, and feedback loops that make models improve week over week. If you want ownership across research, engineering, and deployment, and you want to see your work move real business metrics for merchants, Gentoo may be the place for you.
What you’ll be doing
•
Design, build, and operate multi agent systems that understand user intent and drive the right actions in real time.
•
Own retrieval augmented workflows end to end, from embeddings and indexing to ranking and feedback loops that continuously improve recommendations.
•
Define agent orchestration, memory, and context handling, and build the data pipelines that keep information flowing from collection to labeling to evaluation.
•
Improve evaluation systems, including LLM as a judge, run offline metrics and online experiments, and use the results to decide what ships.
Qualifications
•
Strong foundations in machine learning and applied NLP with a record of shipping production systems
•
Proficiency in Python and the ability to turn research ideas into maintainable, testable code.
•
Experience with retrieval augmented systems, including prompt and policy design and integration with vector search.
•
Ability to design clear evaluation metrics and run offline evaluation or A/B tests to guide launches.
•
Product mindset with attention to latency, cost, and the user experience behind every response.
•
Ownership, sound judgment, and clear communication in a fast moving, ambiguous environment
•
Our current stack: Python, FastAPI, Airflow, AWS, Vercel, Pinecone, plus TypeScript and React for surfaces and NestJS for services. Familiarity is a plus, not required.
