Work
Sanitized case studies from client-facing AI, data, and platform engineering work, plus earlier full-stack projects. Each case study covers the problem, my role, the approach, the tradeoffs, and the outcome.
Case studies
An LLM layer on top of existing Snowflake forecasts that explains, in plain language, how external events affect confidence in fuel-consumption forecasts.
Adopted to help stakeholders interpret forecast confidence and external drivers.
A from-scratch replication engine that detects change on source systems with no native CDC, using hierarchical hashing to land Fivetran-format SCD2 history in Snowflake.
Validated end to end; production rollout was in progress when I left.
Hardened an AI governance framework for Snowflake semantic views and Cortex agents: a CI drift guard, a Snowsight-to-git capture step, and a human approval gate before irreversible production deploys.
Framework adopted by multiple organisations; hardening changes merged upstream.
A Streamlit and Snowflake operational control board with near real-time refresh, built through iterative consultation with designers, analysts, and engineers, and adopted for operations alignment.
In production, providing a shared operational view across three groups.
A production reporting workflow that generates first-pass explanations of monthly KPI movements while analysts retain responsibility for validation.
Integrated into monthly production reporting, reducing first-pass manual discovery.
Archive
Secure veterinary platform with appointment booking, medical records, prescription refills, and an admin dashboard over authenticated workflows.
WebSocket messaging with JWT authentication, VPS deployment, and CI/CD automation; the concept informed client communication workflow design.
Full-stack grocery web application with product browsing and authenticated user workflows.
Algorithmic maze generation and solving, extending a Minecraft C++ API.