Projects

A selection of projects I've worked on.

Want a paper copy of my CV?

Digital Apprentice Service

Maintaining and developing the infrastructure for all services under the Digital Apprentice Service, leveraging Azure Cloud and Infrastructure as Code best practices. Focused on optimizing deployments, modernizing pipelines, and ensuring robust, scalable cloud environments.

DevOps Engineer (Jul 2023 – Present)

Achievements

  • Developed new PowerShell scripts and integrated them into CI/CD pipelines to automate Azure resource deployments
  • Converted legacy classic pipelines to modern YAML-based pipelines for improved maintainability and transparency
  • Led the migration from ARM templates to Bicep, streamlining infrastructure-as-code and enabling easier resource management

Technologies

AzureAzure DevOpsARMBicepPowerShell

Home Office

Core engineer for Cerberus, a Spring Boot-based microservice with a React frontend that enabled internal border agents to view, filter, and analyze consignments and air packages entering the UK. Cerberus supports national security by surfacing critical data at the right time for decision-makers.

Full-Stack Engineer (Nov 2020 – May 2023)

Achievements

  • Designed and developed a custom UI library from scratch to standardize and accelerate front-end development
  • Implemented advanced sorting and filtering mechanisms to streamline user workflows and surface the most relevant consignments
  • Built an automated reporting system using Jasper Reports, delivering actionable reports to end-users monthly
  • Transformed third-party vendor data from Kafka topics using complex business logic to ensure clarity and usability within Cerberus

Technologies

TypeScriptJavaSQLSpring BootReduxReactKafkaJasper Reports

HM Revenue and Customs

Engineered EDHMT, an entity matching tool used to analyze HMRC’s monthly data runs for fraud detection. Integrated with BAE's NetReveal software to enhance entity matching and support the MCR team’s advanced data analysis workflows. The system helped HMRC uncover complex tax fraud patterns across vast data sets.

Software/Data Engineer (Aug 2018 – Sep 2020)

Achievements

  • Developed custom algorithms for precise entity matching to support tax fraud investigations
  • Deployed the solution to distributed clusters, optimizing performance on edge nodes
  • Integrated monthly HMRC data with NetReveal, automating analysis and reporting for the MCR team
  • Enabled scalable and high-throughput processing for large-scale government data sets

Technologies

JavaDistributed Computing

furqanagwanPrivate

My personal website and digital portfolio, fully designed and developed from scratch. Features project case studies, a technical blog, and structured MDX-based content—powered by a modern Next.js stack and a custom design system. Content is sourced directly from a private GitHub MDX repository for maximum flexibility and version control.

Creator, Designer & Full-Stack Engineer (2018 – Present)

Achievements

  • Designed a custom UI/UX system and implemented responsive layouts
  • Integrated dynamic MDX content rendering with category filtering and rich markdown features
  • Connected to a private GitHub content repository via API for seamless, version-controlled content updates
  • Deployed globally on Vercel with statically generated pages for speed and SEO

Technologies

ReactNext.jsTypeScriptTailwindCSSMDXVercel

furqanagwan-contentPrivate

A private GitHub repository acting as a headless CMS for the furqanagwan website. Stores all MDX files and static images for posts, recipes, and project documentation. Enables simple, version-controlled content updates and easy collaboration via standard GitHub workflows.

Creator & Maintainer (2018 – Present)

Achievements

  • Structured content storage using MDX and folders for posts, images, and project data
  • Established secure, token-based API access from the website to fetch and render content at build time
  • Enabled version control, history, and collaborative editing of all website content using standard GitHub processes

Technologies

MDX

metabolicai Public

A modern, interpretable AI-powered TDEE and calorie analytics app. Users log daily weight and calorie data, and the backend uses XGBoost (with lagged time series features) to predict real, personalized TDEE values. Built with FastAPI, SQLite, and a CI/CD pipeline for robust, test-driven deployment. Features real-time retraining, advanced feature importance analytics, and full test coverage.

Creator, Designer & AI Engineer (2025 – Present)

Achievements

  • Designed and implemented a FastAPI backend with SQLite and full CRUD endpoints for health tracking
  • Developed an XGBoost-based model with custom feature engineering (lag, moving average, demographic features) for personalized TDEE prediction
  • Automated CI/CD pipeline with >90% code coverage using pytest and GitHub Actions
  • Surfaced model feature importances to users, enabling transparency and actionable insights
  • Implemented robust test cases for all API branches, reaching near-100% code coverage
  • Engineered retraining logic to auto-update models on every new user entry

Technologies

PythonFastAPISQLiteXGBoostscikit-learnPytestGitHub Actions