AI & Cloud FinOps Engineer

WTW is evolving the FinOps practice that keeps our AI and cloud spend visible, accountable and efficient.

This role will help drive the evolution from traditional cloud FinOps to Tokenomics, and the automation of that practice.

You will build the controls, reporting, and automation that let our FinOps practice run without manual effort across Azure and AI, and you will help stand up our FinOps for AI capability from a standing start.

Small core FinOps team, global estate, and work that is visible to our most senior technology leaders.

We focus on 4 core areas: how we plan spend, how we monitor it, how we optimize it, and how the practice itself evolves.

Plan: design cost in before it is spent

Automate design-time cost estimates, so the cost of a workload or an AI feature is understood before it is built.

Support continuous forecasting by support the automation of the monthly forecast and variance pack, including AI-drafted commentary, published to stakeholders each month.

Monitor: make cost visible without manual effort

Automate the ingestion, normalization and allocation of cloud and AI cost data, aligned to FOCUS, so that reporting maintains itself instead of being rebuilt each month.

Instrument AI cost attribution at source; Team, product, feature and environment on every call so that token and model consumption can be attributed, shown back and charged back.

Build the alert-to-closure workflow: multi-threshold budget and anomaly alerts routed to a named owner, with response SLAs, escalation, and closure evidence captured automatically.

Enhance monitoring with automated metrics that lead to meaningful and actionable insights.

Optimize: turn insights into action

Automate insight-to-action routing so optimization recommendations, anomalies, and trends reach the right owner, are tracked to resolution, and do not quietly expire.

Automate cost hygiene: detection of orphaned and idle resources triggering lifecycle actions and build the tooling behind our rightsizing practice.

Help shape the guardrails that stop AI spend running away: real-time visibility, cost controls, step and budget caps on agentic workloads, and a central model-routing layer that enforces model-tier policy, caching and batching without teams having to change their code.

Evolve: grow the practice, especially for AI

Build out our FinOps for AI practice: unit economics such as cost per prompt, per outcome, per user and per agent run; model, seat and feature hygiene.

Apply AI to the practice itself: anomaly explanation, tagging remediation, variance commentary and optimization prioritization.

Codify and automate best practices, so engineering, product and finance teams can act without queuing behind the central team.