epharma — Tech Manager / AI Engineering Manager
Dec 2025 – present
São Paulo
Head of AI and of the development teams. 7 tech leads, ~60 developers and QAs, Innovation Hub.
- Built an internal AI Factory that cut average delivery lead time from ~35 to 12 days and brought AI-assisted development to 100% of squads.
- Agentic Engineering ecosystem with Claude Code, LiteLLM and Azure AI Foundry covering the full SDLC: story writing and tracking with the business, code generation, PR review and deployment. Adoption sustained by internal skills and a company-owned marketplace.
- Trained the teams on Claude Code and agents across multiple sessions on AI-assisted coding and productivity. The 100% adoption came from training, not from mandate.
- POC governance through Innovation Committees open to the whole company. 63 hypotheses taken to MVP, 6 in production. The funnel exists to kill weak ideas early.
- Legacy handled with observability before decisions: no refactoring on impulse. Same standard applied to the area's CAPEX and OPEX.
- Migration from Azure DevOps to Jira with standardized project metrics; Coolify alongside Kubernetes for experimentation environments.