AI Engineer

About the role

The AI Engineer will work directly under the Head of AI to build and ship the intelligence layer that powers the product. AI is not a feature here; it is the core of how families get instant eligibility decisions, and how the company scales compliance without scaling headcount.

Responsibilities

• Build MVPs from scratch: take new AI products from zero to real users — both consumer-facing and internal tooling — with minimal hand-holding and a high bar for quality
• Optimise accuracy and latency: tune LLM and VLM pipelines, and classical ML models where appropriate, to meet the standards a regulated fintech product demands
• Create robust evals: build evaluation frameworks that make AI behavior measurable, reproducible, and improvable over time — so regressions are caught before users feel them
• Read data and fix mistakes: diagnose real-world AI failures by going directly into the data, making ad-hoc corrections, and closing the loop fast
• Build endpoints and tooling: surface AI capabilities to teammates in reliable, well-documented ways so the whole team can move faster without depending on you for every query
• Work across the full AI stack: primarily LLM and VLM-based in early stages, with scope to fine-tune or train models from scratch as individual products mature and optimisation demands it

Requirements

• Biased toward simplicity: you know that managing many AIs gets complex fast — you resist unnecessary abstraction and build systems that are easy to reason about and maintain
• Values old and new AI equally: you recognise the tradeoffs between prompting, fine-tuning, and training from scratch — and you pick the right tool for the job rather than defaulting to the latest trend
• User-obsessed: AI is the blocker to a good number of AHA moments in the product — you keep the end-user in mind in every technical decision, not just the benchmark
• No task beneath you: reading data, making database edits to correct AI mistakes, writing evals for edge cases — you treat this as essential product work, not a distraction from "real" engineering
• Comfortable with ambiguity: you can scope your own work, define your own quality bar, and ship without waiting to be unblocked

Qualifications

• Able to work in-person with the team in San Francisco, CA or Phoenix, AZ (visa support available)

Skills

• Experience with LLM APIs, vector databases, fine-tuning pipelines, or evaluation frameworks is a strong plus

Pay

$125,000 - $175,000 per year, commensurate with experience

Schedule

In-person