Felipe Cruz

Results

What I deliver

Concrete metrics and scope — not a generic skill list.

• Patented DMEPOS SaaS in production across 350+ US clinics — DME domain (ordering, eligibility, clinical supply workflows) with HIPAA requirements and ISO 27001-aligned technical and organizational controls engineered into the product and operations.
• Fleet of 1,000+ clinical iPads and 200+ corporate endpoints: procurement, OEM-to-employee fulfillment and zero-touch enrollment via Windows Autopilot and Apple Business Manager (ADE); policy and compliance through MDM, identity on Microsoft Entra ID, perimeter hardening with WAF and end-to-end security posture.
• Engineering scale 1 → 15+ ICs; stood up Frontend, Backend and QA as durable practices. SDLC driven by a task system with explicit stages (in progress, in development, review, release), delivery cadence, Definition of Ready/Done, code review and quality gates aligned with product squads and roadmap execution.
• AI-accelerated delivery on a flagship legacy rewrite: in a few months, replaced a ~7-year-old production system with a leaner, higher-quality codebase; moved the full engineering org onto the new platform and kept shipping with AI-assisted development. On the order of ~500% gain in team productivity, a sharp reduction in bugs and legacy performance issues, and clearing the bulk of technical debt in that same window using leading code-centric models.
• Greenfield Brazil subsidiary for a US remote-first parent: registered address, physical office, business-grade connectivity, voice and collaboration stack; fit-out, access control and physical security. Full-cycle hiring (sourcing through onboarding) for 70+ roles across functions; ~$6M cumulative efficiency impact and ~95% retention over a ~7-year window.
• ELT analytics platform: Fivetran → Snowflake → dbt → Apache Superset — ingestion, versioned transformations and a semantic layer; governed self-service analytics. Snowflake-native AI agents / analytical copilot so strategy consumes insights and curated reporting without ad hoc SQL; RAG, LLM orchestration and MCP where they strengthen the product surface.
• Full-stack observability on Datadog: APM with distributed tracing (correlated spans, latency waterfalls), structured logging, Database Monitoring on critical workloads and infrastructure metrics. Slack-routed alerts with severity and actionable context; triage discipline to drive down MTTR and catch regressions before the first customer ticket.
• High-concurrency OLTP database optimization at scale: execution-plan analysis, predicate-aligned indexing, materialized views where appropriate, N+1 remediation and disciplined ORM usage. Roughly ~80% reduction in critical-path database round-trip latency and ~90% on target query duration, measured on production traffic using P90 and P99 percentiles.