AI Engineer

YipitData is the leading market research and analytics firm for the disruptive economy, and they are seeking an AI Engineer to turn billions of alternative-data points into actionable insights. The role involves applying data science with AI-native tooling to lead analytical projects and collaborate with various teams to enhance customer-facing products.

Responsibilities:

• Translate ambiguous customer questions into well-scoped data science projects spanning panel data, time series, and causal inference
• Engineer features from large alternative-data panels (transaction-level, invoice-level, web-scraped) using Spark
• Build, validate, and interpret causal and predictive models that link alternative-data signals to financial outcomes such as revenue, earnings surprise, and KPI inflections
• Author technical white papers and customer analyses for institutional investors and Fortune 500 readers, including figures, equations, and narrative framing for sophisticated readers
• Use LLM coding assistants and agents as a primary collaborator: prototype faster, write higher-quality code, audit your own work, and ship deliverables in compressed timelines
• Build internal LLM-driven tooling (agents, eval harnesses, retrieval pipelines) for the broader organization
• Partner with data engineering, product, and revenue teams to close the loop between signal development and the customer-facing product
• Set technical standards for the team: PEP 8, type hints, vectorized operations, reproducible notebooks, sound methodology, and citation discipline

Requirements:

• You have shipped multiple data science projects end-to-end and can point to quantifiable customer or business impact
• Your statistical foundations are real, not surface-level. You can defend a causal-inference method choice end-to-end under technical questioning
• You write Python that reads like the work of a senior engineer: clear naming, type hints, vectorized code, no premature abstraction
• You already use LLM coding assistants daily and can describe specific cases where they multiplied your output without compromising quality, and specific cases where you overrode them
• Writing is a primary skill for you, not a side hobby. You communicate as carefully on the page as you do in code
• You take ownership without prompting: you find the right question, ship the answer, and explain it
• Humility about what you do not know matters as much as confidence about what you do
• You read the literature, cite sources, and ground your work in established methods rather than reinventing them