FinancialControl — CSV Ingestion & Rules Engine
A financial-operations MVP that ingests inconsistent CSV exports, normalises them, and runs a configurable rules engine to classify, flag, and route transactions automatically — with a clean review UI and a full audit trail.
FinancialControl is a financial-operations platform that turns messy, multi-source CSV exports into clean, categorised, auditable transaction data. A pluggable parsing layer absorbs the inconsistencies of real-world exports, and a configurable rules engine — conditions mapped to actions — classifies and routes each record deterministically, so the same input always produces the same result.
Built as a .NET solution with an ASP.NET Core API, PostgreSQL persistence via Entity Framework Core, and a Docker deployment. A review interface surfaces flagged and uncertain items for the owner to confirm or override, and every change is recorded. Delivered over six milestones and now the backbone of a wider suite of internal systems for the client's business.
Case Study
The Problem
Motti's company was reconciling financial data by hand across several systems, each producing CSV exports in different shapes, encodings, and column orders. Categorisation logic lived in people's heads, there was no consistent way to flag anomalies, and nothing left an audit trail. The process was slow, error-prone, and impossible to scale as the business added more revenue streams and brands.
The Approach
Built a .NET backend (ASP.NET Core, PostgreSQL via Entity Framework Core, Dockerised) around two pieces: a pluggable ingestion layer that tolerates messy real-world CSVs — inconsistent headers, encodings, and delimiters — and a deterministic, client-configurable rules engine that maps conditions to actions to categorise and route each transaction. Added normalisation, de-duplication, and a review UI where the owner confirms or overrides flagged items. Shipped across six milestones so the client could validate the logic incrementally before committing.
The Outcome
Delivered and in daily production use, with the owner reconciling automatically instead of by hand. It became the first of several systems built for the company — Task Dispatcher, the Villa decision layer, and a marketing agent all followed. Full source and documentation handed over; every milestone released and paid.
“The best freelancer I ever work with! With his help, I am building several support and management systems for my company. Helgi is a high-quality and diligent professional and identifies my needs before I even recognize them.”
— Motti Turgeman