AI Engineer (Consultant)

Overview

As a Finance AI Engineer at Accenture, you design, build, and operate production-grade AI applications that automate finance processes and enable trusted decision-making. You work within a delivery pod on client data, translating finance requirements into secure, observable software and reusable patterns. You’ll craft agent-based workflows, integrate with ERP/EPM systems, and ensure measurable quality, latency, and cost. This role combines deep engineering with finance-domain insight to deliver high-impact, scalable results.

Responsibilities

• Build and deploy agent-driven workflows in live finance processes (reconciliations, journals, exceptions, disputes, collections, variance explanation) with clear acceptance criteria and controlled failure behaviour
• Implement LLM application patterns (tool calling, structured outputs, retrieval, multi-agent orchestration, versioning)
• Integrate with ERP/EPM systems, document repositories and workflow tools; manage data extraction, transformation and secure APIs
• Write evaluations, instrument behavior, and ensure visibility of failures using automated tests, observability, red-team scenarios, and metrics for quality, latency, and cost
• Develop finance user interfaces for reviewing, approving, overriding, and evidencing agent actions with provenance and accessible error recovery
• Package and operate solutions using cloud AI services, containers/serverless, CI/CD, IaC, secrets management, monitoring, release controls and rollback
• Collaborate with finance users, process experts, architects and data engineers to clarify requirements, document runbooks and operating procedures
• Contribute reusable patterns and components to the practice asset base (code, evaluation assets, reference implementations, guidance)

Key requirements

• Strong software engineering discipline (Python and at least one other production language e.g., TypeScript)
• SQL, API design, code review, secure coding practices
• Hands-on LLM application development (tool use, structured outputs, retrieval, agent frameworks)
• Experience with evaluation, observability and LLMOps/AgentOps (versioning, monitoring, release, rollback)
• Understanding of secure enterprise integration and sensitive-data handling (authentication, authorization, secrets, logging, PII-aware design)
• Ability to work with non-technical finance users translating needs into testable requirements
• At least 4 years’ professional experience

• Effective collaboration with non-technical stakeholders
• Clear communication of options, risks and trade-offs
• Constructive challenge of assumptions
• Python
• TypeScript or other production language
• SQL

…