PLUS8SOFT reposted this
We built an AI that detects money laundering. And we're giving it away for free. Three years ago, when we started building software for fintech clients, I kept hearing the same thing from compliance teams: "Our current system flags 50,000 alerts a month. My analysts review 2% of them. We know we're missing real cases - we just don't have time." That stuck with me. Financial crime drains trillions from the global economy every year. Yet most mid-market banks and fintechs still can't access enterprise-grade AML technology. The cost is brutal, integration drags on for a year, and that's after the 6-month sales cycle. So we built something different. Introducing PLUS8SOFT AML Detection Pipeline (Community Edition). A production-ready Python ML pipeline for detecting suspicious financial transactions. Free. Open source. Apache 2.0. We spent 6 months just on the graph module. Early versions kept missing layered transactions through shell accounts. It took several iterations to catch them reliably. What's inside: ▸ Graph-based transaction analysis (NetworkX) to surface hidden network patterns ▸ Correlation-driven feature engineering to cut noise ▸ Three ML models out of the box: RandomForest, XGBoost, LightGBM ▸ Automatic best-model selection based on your data ▸ F1 score of 0.92 on benchmark datasets (95% recall) For teams that need more - production SLAs, custom integrations, dedicated support - we offer an Enterprise Edition. This is the tool I wish we had when we started. If it saves one compliance team from triaging 50,000 false alerts a month, it was worth building. Compliance and risk teams - I want to hear from you: What's the biggest gap in your AML stack right now: false positives, integration time, cost, or something else? Drop your #1 pain point in the comments. We're prioritizing the next release based on what we hear this week. ⭐ Star the repo if useful - link in the first comment. Enterprise: sales@plus8soft.com #AML #FinCrime #RegTech #OpenSource #FinTech #Compliance