I have experience usying of different Machine Learning and Deep Learning tools (PySpark, Scikit-learn, Tensor Flow, NLTK, Pandas among others) to structure collection campaigns
Using neural networks (Tensor Flow in Python), the process of homologation of vehicle brands and lines was automated for the settlement of taxes, reducing the operational load by 75%.
With neural networks (Tensor Flow in Python) and collection campaigns were structured for operations, managing to increase collection between 21% and 33%.
With Python and Pandas, a virtual Supervisor was built for operations that, by identifying the Out Layers (atypical cases), generates alerts to prevent the expiration or prescription of the traffic infraction processes, increasing collection and avoiding fines.
With Phyton, Pandas and Tensor Flow, the customer acquisition campaigns with the highest probability of renewal were structured from tables of 15 million records for the Technical Mechanical Review and SOAT partners.
With Python, Pandas and databases in PosgreSQL, the process of reporting allies, reconciliation and invoicing was automated.
With PySpark, Random forest classifier and NLTK (Natural Language Toolkit) a model was developed to predict with 82% accuracy the candidates who will pass the first stage in a selection process (Hackathon).
Work Terms
Preferably payment per deliverable