CDR Dispute Matching Tool

I’m looking for a developer who can build a lightweight tool that connects directly to our telecom databases—specifically the raw CDR logs—and automatically matches records to flag potential billing disputes.

The workflow I have in mind is simple: pull daily CDR dumps from our switch, compare them against the reference set we receive from our interconnect partners, and highlight every variance in duration, destination, or missing record so we can escalate disputes immediately instead of at the end of the month. Speed is critical; we handle millions of rows a day, so the matching engine must be memory-efficient and able to stream or batch-process large CSVs without choking.

Preferred stack is entirely up to you: Python with pandas, Spark, Go, even a compiled C++ utility—whatever lets you parse, match, and output results quickly. A small CLI or web dashboard for uploading files and exporting a clean discrepancy report (CSV or JSON) would round things out nicely, but the matching accuracy and performance take priority.

Deliverables
• Source code for the matching engine and a brief setup guide
• Configurable matching rules (time tolerance, CLI/CLD fields, duration thresholds)
• A sample report that clearly lists all disputed records with reason codes
• Short documentation covering installation, usage, and how to add more CDR formats down the line

I can provide anonymised CDR samples for initial testing as soon as we get started, and I’ll be available for quick feedback loops so we can fine-tune the detection logic together. Let’s build something that saves both of us hours of manual reconciliation every billing cycle.