James Darby builds data platforms and AI systems.

• Cloud cost — Cut cloud spend by 50% on my trial day, with a roadmap to a 90% total reduction.
• Data platform — Own it end to end, from architecture to delivery: cost-efficient ETLs feeding client-facing dashboards and data-sharing APIs.
• AI insights — Building features that turn our data into actionable insights delivered directly to customers.
• Infrastructure — Architect cloud infrastructure and security across the platform.

• AI alert-review agent — On the agentic team, built an agent adjudicating screening alerts: 2.5 years of human analyst work completed in 12.5 hours. On a labelled dataset it wrongly dismissed 5 of 1,000 true positives — 99.5% against a 99% safety bar, beating human error rates — with a second verification agent as backstop.
• Multi-tenant lakehouse — Core engineer on a multi-cloud (AWS + GCP) Databricks Delta Lakehouse powering live client dashboards and exports. Kotlin Spark Structured Streaming ingestion of Kafka Protobuf events with per-tenant KMS encryption, run in cost-saving batch mode: ETL time from 12 hours to 20 minutes.
• Config-driven ETL — Co-built a PySpark framework with declarative column maps, data-quality checks, dedup and latest-state/history merges — new pipelines ship as YAML, not code. CI diffs view changes against production data to catch breaking changes and perf regressions before release.
• Multi-tenant BI — Embedded Superset with dashboards-as-code and row-level-security dynamic views: one dashboard and one table serve every tenant, replacing per-client duplication.
• Front end — React interface with Cypress-Gherkin E2E tests, localisation, component libraries and Vite/NX package splitting: CI/CD from 60 minutes to 15.
• Microservices — Kotlin-Spring services over Yugabyte, Databricks and Superset with sub-15-second loads; canary and blue-green releases via ArgoCD, observability with Datadog, Sentry and Grafana in a SOC 2 environment.

• Full-stack architecture — Designed a complete overhaul from DevOps to front end in Terraform, React TypeScript, Django and AWS.
• Migration — Managed the migration of 6 million customers to a new system.
• Leadership — Planned hiring, trained new engineers and led teams through technical design and architectural planning.

TypeScript, Python, Kotlin, SQL (Postgres), Terraform, HTML/CSS

React, Django, Spring, Spark, Pytest

Git, Docker, AWS, GCP, Databricks, Superset

Zustand, React Query, RTL, Styled-Components, MSW, Pandas, NumPy

Software Engineering Immersive · 2019 — 2020

BA Product Design · 2014 — 2017