Posts

Showing posts from August 13, 2026

I Built a BFIU-Compliant AML Detection System in Python (Here's Why the Kaggle Approach Doesn't Work

I Built a BFIU-Compliant AML Detection System in Python (Here's Why the Kaggle Approach Doesn't Work)

-
Most AML tutorials end with a confusion matrix and a 99% accuracy score. Here's why that doesn't work — and what I built instead. I've been working in fintech compliance data for a while. The one thing I kept noticing: every "fraud detection project" on GitHub or Kaggle uses the same dataset — the UCI credit card fraud dataset from 2013. It has 284,000 rows, 30 features labeled V1-V28, and approximately zero explanatory value for anyone who wants to understand how financial crime actually works. So I built something different. The problem with the standard approach Real transaction monitoring engines don't work like Kaggle competitions. They don't take a CSV, train a model, and output a probability score. They work like this: A rule engine runs first — deterministic, auditable, regulatory-cited rules that generate alerts Those alerts get scored and triaged by risk tier An ML layer reduces false positives among the high-risk alerts ...
Post a Comment
Read more

8 Years of Hunting Anomalies: How Isolation Forest and Autoencoders Changed My AML Game

-
I still remember the day our team detected a massive structuring ring, involving over 500 fake accounts and BDT 50 million in suspicious transactions. It was a high-stakes scenario - if we didn't report it to the BFIU within 24 hours, our MFS license would be at risk. The Hidden Problem Standard machine learning approaches often fail in Bangladesh due to the unique characteristics of our transaction data. With over 100 million mobile financial service (MFS) users, the sheer volume of data is overwhelming. Moreover, the BDT 100,000 threshold monitoring and STR/SAR bottlenecks make it challenging to identify true anomalies. That's where Isolation Forest and Autoencoders come into play. Both algorithms have their strengths and weaknesses, but when combined, they can be a powerful tool in identifying transaction anomalies. Technical Breakdown & Logic Flow Isolation Forest is an unsupervised learning algorithm that identifies anomalies by isolating them from the rest of the da...
Post a Comment
Read more