AI Engineer

Role overview

Build and own the scoring and signal layer of a data platform that combines commercial datasets, public web information, and team notes into ranked, explainable company insights. You will shape how scores are calculated, validate their quality, and improve them with regular feedback from business users.

Responsibilities - Combine records from varied sources, resolve entity matches and conflicts, and turn the resulting data into usable signals. - Develop and maintain ranking, scoring, or prioritization methods, explaining their outputs to business stakeholders. - Build production LLM workflows such as agents, retrieval-augmented generation, or structured extraction. - Create evaluation sets and regression checks to measure output quality and assess changes to prompts or systems. - Use user feedback and signal trends to identify stale or unhelpful rankings and guide improvements. - Define business metrics and work directly with leadership and users to review scoring decisions.

Requirements - At least five years in data science, analytics, or machine learning, including three or more years building production LLM systems. - Strong Python and SQL skills for querying data and developing features independently. - Experience delivering scoring, ranking, propensity, or lead prioritization systems and defending their results to stakeholders. - Practical LLM evaluation experience, including quality measurement and regression testing. - Experience reconciling messy, conflicting, multi-source data, including entity matching. - Familiarity with feature stores, vector databases, or embedding-based similarity search for company or people data. - Excellent spoken English and confidence discussing metrics with non-technical colleagues. - Valid passport and current US B1/B2 visa.

Nice to have - Private markets data experience, including funding, revenue estimates, ownership, or company relationships. - Knowledge of commercial data provider APIs and their coverage, freshness, rate-limit, or identifier challenges. - Time-series or cohort analysis of company-level signals; experience with investment or dealmaking teams; or human feedback systems for improving models. - Experience shipping agentic systems, contributing to LLM evaluation or scoring work, or using AI-assisted development workflows.

Benefits and work setup Remote from Latin America, with preference for Costa Rica or Central America. Travel to New York is required for two to three weeks at kickoff and approximately every other month afterward.