Stack

The research layer always comes first. I/O Psychology frameworks, validated psychometric instruments, and statistical modeling define what gets built. Python and R handle the analysis — the algorithms emerge from the science, not the other way around.

AI-augmented architecture with Windsurf as the primary IDE agent for multi-file work and codebase indexing. Next.js and TypeScript on the frontend. Netlify for edge deployment.

Serverless by default. Firebase and Firestore for production data, Cloud Functions for scheduled and event-driven logic, Secret Manager for credential hygiene.

DeepSeek V4 as the daily driver for content generation and scripting — cost-efficient and capable. Claude for architecture decisions. Google Colab for cloud-based Python execution without local environment friction.

Enterprise-grade visibility. A decade of building analytics systems inside financial institutions — Power BI and Tableau for executive-facing dashboards, SQL for the architecture underneath them.

Large-scale GRC platform management, low-code automation, and risk workflow architecture across regulated financial environments. The past-tense part of the stack — and the source of the pattern recognition I apply everywhere else.