Engineering manager who builds.
I came up through tax audit and taught myself to code. Spent the last 6 years at Deloitte — leading a team of engineers for the last 3 years while owning a portfolio of production-grade internal and external applications.
What sets me apart is full-stack engineering fluency and the ability to carry a fuzzy business problem through to a live product — requirements, architecture, and delivery, without handing off at any stage.
Outside work, I build and maintain open-source software keeping me close to the tooling and infrastructure I bring into the day job.
Lead a small team of engineers (typically around four, varying by project and year) building and maintaining internal applications for Deloitte's tax, legal, and finance departments — used daily by colleagues across those same departments. Own the complete product lifecycle: stakeholder interviews and requirements, system design, hands-on development, cloud deployment, and ongoing iteration. Also carry all project management for the team.
Tax authority reconciliation engine
Reconciles large-volume tax data against internal records. 100+GB MongoDB database; each run retrieves and processes GBs of data. FastAPI on Azure Web Apps within the firm's private VNet / hub-spoke network.
e-Delivery Note Tool
Automates reconciliation between warehouse records and myDATA transmissions, replacing manual, exhausting checks with continuous monitoring. Surfaces discrepancies and helps ensure 100% accuracy and compliance. Built for Deloitte's Tax & Legal team.
Document intelligence platform
Started as an invoice reader; evolved into a general-purpose document processing platform. Uses Azure Document Intelligence and an LLM layer for extraction and reasoning. Django, deployed on Azure.
Contract management system
Centralised repository for the firm's contracts with full-text search and lifecycle tracking. Replaces a fragmented manual process. Django, deployed on Azure.
Fleet & benefits management app
Used by the finance department to manage employee vehicle benefits and associated reporting. Django, deployed on Azure.
No-code automation server
Allows non-technical staff to trigger and configure automated workflows without writing code. Detects available automation parameters and surfaces them in a UI for user-driven execution. Ruby deployed on AWS, Django deployed on Azure.
Python (FastAPI, Django, Flask, Pydantic) · MongoDB · PostgreSQL · Azure Web Apps, VNets, hub-spoke networking · Azure Document Intelligence · OpenAI API / open-source LLMs (Hugging Face) · Docker · GitHub Actions · Azure DevOps · Ruby · Angular
Tax audit for corporate clients. Built the finance and regulatory domain knowledge that now directly underpins the products above — and clarified exactly where automation could remove the most risk and manual effort.
Crati
Platform for processing and analyzing Greek government transparency documents from Diavgeia — ingests decisions, extracts text, maps entities and relationships, and exposes search, exploration, and monitoring tools. Built on the diavgeia-api client below.
Django · Celery · PostgreSQL + pgvector · Redis · RabbitMQ · OpenSearch · Nginx · Docker · Grafana / Jaeger / Loki · AWS S3 · Coolify
diavgeia-api
Python client for the Diavgeia Open Data API — typed Pydantic models and simple / advanced search. Published on PyPI, MIT licensed.
Also published on PyPI: scrape-reviews · secdata