Benjamin Tovar Cisneros

Senior Data Scientist

Credijusto.com•  January 2019 - March 2021

Credijusto.com is a Mexican Small Medium Enterprises [SMEs] lending FinTech/Neo-bank, LinkedIn Top 2020 startups - 3rd place. Job description: - Responsible of developing end-to-end Machine Learning models. Starting from scoping business needs, framing the business question(s), collect, clean, analyze, test hypothesis to ensure project viability, and model implementation in production, followed by model updates and monitoring. - Responsible of meeting stakeholders to scope project requirements and coordinate with other teams such as product and engineering. - Responsible of ensuring that our data pipelines are properly designed to serve different clients, such as core data analytics to risk analyst, to marketing and operations. - Decision science tech lead, responsible of planning and managing the team of decision scientist. - Responsible of interviewing and supporting the hiring process of data scientist/analyst and data engineering. Results: - Reduced manual time collecting historical data from our clients from 3 hours average to minutes by building an ETL in Python that unified different internal credit memo versions and made the data available and consistent as a MySQL DB available for all Data Scientists / Analysts [Python (Pandas, Numpy), Apache Airflow, MySQL, git]. - Reduced credit analyst time spent on interviews and manual document review for restaurant applications from an average 6 hours to minutes by implementing multiple regression models that automatically estimated cashflows by combining bank statements, electronic invoices, credit reports and geo-sociodemographic information that corrected income tax underreporting, which is a common problem in Mexico [Python (Scikit-learn, Pandas, Numpy), MySQL, Jupyter Lab, MS Excel, git]. - Reduced credit analysts time spent through manual review of electronic tax declarations to real time standardized financials KPI metrics from an average 3.6 hours to minutes. As the Decision Science tech lead I helped to coordinate the design and implementation of an end-to-end data pipeline to extract, process and calculate credit risk metrics for millions of electronic invoices [Python (Pandas), AWS EC2/S3, PostgreSQL, Jupyter Lab, Lucidchart].

Senior Data Scientist

Alba AI•  August 2017 - December 2018

Alba AI is a Consultancy E-Learning startup, providing Recommendation Systems to online & on-site education platforms in Mexico. Job description: - Responsible of developing end-to-end Machine Learning models. Starting from scoping business needs, framing the business question(s), collect, clean, analyze, test hypothesis to ensure project viability, and model implementation in production, followed by model updates and monitoring. - Responsible of meeting stakeholders to scope project requirements and coordinate with other teams such as product and engineering. - Responsible of ensuring that our data pipelines are properly designed to serve different clients, such as core data analytics to risk analyst, to marketing and operations. - Responsible of interviewing and supporting the hiring process of data scientist/analyst and data engineering.
 Results: - Implemented an end-to-end Recommendation System (RS) that improved educational minigame success rate (success/failure) to 80% +/- 20% from a baseline of 42% +/- 41% by assisting children to navigate through recommended minigame routes based on their topic preferences and managing game difficulty in real time. RS model improved exploration rates (playing different minigames) by 33% and reduced game repeatability by 11% [R (ggplot2, dplyr), Python (Scikit-learn, Pandas, Numpy), PostgreSQL, AWS EC2, Jupyter Notebooks, git]. - Implemented an end-to-end RS model that improved exploration rate (taking different courses) by 82% for Mexican government subsidized self-improvement schools in marginated areas [R (ggplot2, dplyr), Python (Scikit-learn, Pandas, Numpy), PostgreSQL, AWS EC2, Jupyter Notebooks, git].

Senior Data Scientist

Konfio•  June 2015 - July 2017

Konfio is a Mexican Small Medium Enterprises [SMEs] lending FinTech, LinkedIn Top 2020 startups - 4th place Job description: - Responsible of developing end-to-end Machine Learning models. Starting from scoping business needs, framing the business question(s), collect, clean, analyze, test hypothesis to ensure project viability, and model implementation in production, followed by model updates and monitoring. - Responsible of meeting stakeholders to scope project requirements and coordinate with other teams such as product and engineering. - Responsible of ensuring that our data pipelines are properly designed to serve different clients, such as core data analytics to risk analyst, to marketing and operations. - Decision science tech lead, responsible of planning and managing the team of decision scientist. - Responsible of interviewing and supporting the hiring process of data scientist/analyst and data engineering. Results: - Promoted as a Sr. Data Scientist after 11 months and became the team lead. - Implemented an end-to-end Recommendation System (RS) that recommended credit lines to SMEs by linking their credit report profiles such as payment behavior and tax declaration data to optimize payment capacity. 83% of clients accepted the automatic recommended offers after submitting credit applications [R (ggplot2, dplyr, caret), Python (Scikit-learn, Pandas, Numpy), MySQL, AWS Lambda/EC2, Jupyter Notebooks, Linux, git]. - Implemented an automatic fraud detection algorithm that improved default rates by 84%. The algorithm made extensive searches within credit records, identity theft and document forgery [R (ggplot2, dplyr, caret), Python (Scikit-learn, Pandas, Numpy), MySQL, NLP, Image Processing, Tesseract OCR, AWS Lambda/S3, Jupyter Notebooks, Linux, git].

Instituto Tecnológico y de Estudios Superiores de Monterrey (ITESM), Monterrey

Computer Science, MS•  January 2013 - December 2014

GPA: 3.73/4.00 Modules included: Bioinformatics, Computer Science fundamentals, Connectionist and Evolutionary Systems & Research and Innovation Methods. Awards: Academic excellence scholarship holder.

Universidad Autónoma de Nuevo León

Biotechnology, BS•  January 2008 - December 2012

GPA: 3.70 /4.00