Accenture

I work in Accenture’s Responsible AI practice in Latin America, combining AI engineering, agent evaluation, red teaming and technical advisory across enterprise engagements.

Agent evaluation and tooling

I have contributed to evaluation and red-teaming engagements in automotive, telecommunications and financial services. My work includes scenario design, adversarial prompting, multi-turn testing, automated and human evaluation, and failure analysis to guide remediation.

I also build reusable evaluation tools that connect these activities with model-call observability and review workflows. My engineering and evaluation experience spans AWS, Azure and GCP.

Customer-service agent evaluation

I have contributed to several customer-service agent evaluations, including evaluation design, failure analysis and remediation recommendations. We tested responses to inappropriate prompts, identified recurring failure patterns and used the findings to guide engineering changes. In one case, post-remediation testing showed an increase in expected responses from 82% to 90%, with 45% fewer test cases requiring further review.

GAIA · AbbVie

I contributed as an AI engineer to GAIA, an agentic platform combining a knowledge layer with regulatory document generation.

My work on the API and SDK included capabilities for self-service configuration of templates and document types, allowing teams beyond the central platform team to define their own document workflows. I also developed Claude Skills to help teams work with the SDK.

Public project overview — Accenture at AWS re:Invent

Google · AI Security Advisor

Selected as an AI Security Advisor for Google through Accenture.

Solution design

I work with clients and engineering teams to define technical approaches, scope, effort estimates and deliverables for enterprise AI evaluation engagements. I also contribute to Responsible AI governance frameworks and facilitate internal workshops on agentic AI.