Software Engineer, AI FinOps

Role overview

This role builds the financial layer behind a multi-provider AI gateway: estimating per-request cost on each customer's own rates, attributing spend to projects, teams, and workflows, enforcing budgets before requests are sent, and reconciling estimates against what providers actually billed. The work sits at the intersection of data modeling, pricing, and product, where correctness matters more than cleverness.

Responsibilities

- Extend the pricing model to cover negotiated per-provider rates, cache reads and writes, per-unit media pricing, and unpriced traffic - Build attribution and budget features that let a finance lead answer who spent what, on which provider, and why it changed - Integrate provider cost and usage APIs and report the variance between estimate and invoice - Build anomaly detection and forecasting that are explained rather than guessed

Requirements

- 3+ years of backend or data-intensive product engineering - Strong SQL and experience modeling money, including integer units, rounding, and never summing floats - Comfort with TypeScript and reading third-party billing and usage APIs - Care for correctness and for explaining a number to the person who has to trust it

Nice to have

- Experience with cloud cost management, FinOps, or usage-based billing systems - Knowledge of LLM pricing structures across providers - Time-series analytics or a columnar store such as ClickHouse