The Mexican National Banking and Securities Commission (Comisión Nacional Bancaria y de Valores, or CNBV) is partnering with R2A to reengineer its data infrastructure to strengthen its anti-money laundering (AML) supervisory capacity and to accommodate a growing fintech sector in order to:
- Allow financial institutions to submit information for AML compliance digitally and automatically to the CNBV.
- Increase the volume, granularity, and frequency – and improve the quality – of AML-related data submitted to the CNBV.
- Enable CNBV staff to import historical records into the central data storage platform.
- Enable CNBV staff to improve AML-related data validation and analysis, and generate customized reports for supervisory and policy development purposes.
A panel of judges selected Gestell for the development of a data request/storage platform and tools for data-driven metrics and insights for the AML department. Gestell was founded to empower financial institutions and other agencies by utilizing new technologies to create intelligent systems to make financial oversight and operations more efficient and less costly. It has previously worked with the Mexican Notary Offices and various private-sector clients.
In this git repository you will find the following items in both source code and object code format:
| Deliverables | Location in git repository |
|---|---|
| The core Python 3.6 based application with any information and AI packages. | All the *.py files located in the git repository and its subdirectories, third party libraries found in Third-Party-Software text file. |
| The TensorFlow-based AI training and processing. | Located in the MachineLearning directory of the git repository. |
| The Ubuntu server Operating System. | The source for the Ubuntu server 16.04 LTS Operating System can be found in the Third-Pary-Software text file. |
| The Lambda functions for ELT, ETLs, model training and response, Business Analytics and inside movements. | Located in the Lambdas directory of the git repository. |
| The development documentation using the UML framework. | Located in the Documentation directory of the git repository. |
| Git repositories | The repository where this information lives in. (This one) |
| The API consisting of Pipes connected from each of the client’s information locations that will dump the information in a secure way to the database. | Located in the APIs/Layouts subdirectory. |
| The Central Database, a PostgreSQL database where all information will be managed and exploited. | The source for the PosgreSQL Database Management System can be found in the Third-Pary-Software text file. |
| A backlog of every single data entry in accordance with BCBS 239. | The code to generate the backlog can be found in the bcb_tag.py file in the APIs/Layouts subdirectory. |
| Any Pandas logic used to normalize and structure the data, Spark logic (in its PySpark environment) used to manage the data and its processes. | Can be found in the *.py files located in the APIs, Lambdas and MachineLearning directories. |
| The Machine Learning Algorithms, including, but not limited to: Logistic Regression, Clustering (K-means, Affinity and Probabilistic), Neural Networks, and Random Forest Regressors. | Located in the MachineLearning directory. |
| The Data Mining Algorithms that will be used: K-Means, KNN and Apriori. | Located in the MachineLearning directory. |
| Any dashboards and reports dependent on the data, and any patterns discovered while using the tool (all of this will be evolving as more insights are generated), andreactive charts with the ability to download the information. | Located in the Dashboards directory. |
| Big Data Tools, including Data Lake, PySpark and self-learning algorithms. | Located in the MachineLearning directory. |