Avatar

Cesar Conejo Villalobos

Graduate Student/Data Scientist

Biography

Hello! I am Cesar Conejo Villalobos. I have a major in Actuarial Sciences from the University of Costa Rica and a master’s degree in Statistics for Data Science at Carlos III University of Madrid, Spain. Additionally, I have five and a half years of experience in the bank and finance industry. Along this path, I have learned and applied different back-end and front-end data science techniques, especially in data storage infrastructure, data analysis, and machine learning algorithms focused on fraud detection, financial customer segmentation and attrition using SQL and R.

I am also a history and philosophy lover, a passion I try to mix with data science. The most helpful thing that I have learned from history is that it is not much use to make predictions and calculations in the medium and long term. The world always goes on its own business! In the end, things and situations that are unpredictable or very difficult to occur can appear, resulting in the analysis being out of date. So, What can we do in this world of chaos? In my case, I consider that the best thing we can do is quantify the uncertainty with probability. =)

Interests

  • Machine Learning.
  • Statistical Learning.
  • Causal Inference.
  • Operational Research.
  • Reproducible workflows.
  • Data democratization.
  • Coffee.

Education

  • M.S. Statistics for Data Science, 2021

    Carlos III University of Madrid

  • B.S. in Actuarial Sciences, 2014

    University of Costa Rica

Experience

 
 
 
 
 

Consultant, Professional Statistical Analysis

Promidat/FISERV

Sep 2021 – Feb 2022 Costa Rica
I am participating in a product solution for the LATAM section of FISERV. Responsibilities include the design of tables in a Hadoop environment, computation of several KPIs, and dashboard design focused on purchase and payment behavior. The project also provides segmentation analysis of the portfolio of customers, Customer Life Value computation, and prediction models for customer attrition.
 
 
 
 
 

Business Intelligence Senior Analyst

BAC Latam

Sep 2018 – Dec 2019 Costa Rica

I worked on three fronts:

  • Technical leadership: I provided mentorship and technical support for the data science team tasked with detecting the fraud patterns in credit/debit cards and acquiring business. I also reviewed experiments and assisted in the predictions models for other areas of the bank as the Compliance and Credit departments.

  • Information quality: I conducted activities related to ensuring data quality and consistency. I designed interactive dashboards for showing the evolution of the main KPIs and proactively challenge the presence of anomalies and deviations. Dashboards were used for data-driven decisions of the Business team in order to explain historical and current events related to fraud tendencies and chargebacks.

  • Technological solutions: I was responsible for the migration from structured manual business procedures to automatical systems, especially in the Chargebacks department.

 
 
 
 
 

Business Intelligence Senior Analyst

BAC CREDOMATIC

Jul 2017 – Sep 2018 Costa Rica

Responsibilities included:

  • Acquirer commerce review: I created SQL queries and R scripts to detect anomalies and deviations of the daily transactions associated with the distinct business. I used several techniques for detecting fraud cases, as Outlier detection, Association rules, Time series, and other non-parametric approaches.

  • Actuarial Analysis: I applied actuarial validation techniques for the pre-feasibility of new insurance policies for acquirer commerce. Also, I estimated the required technical reserves for this new product based on three different actuarial techniques: Chain Ladder, Buhlmann, and Buhlman-Straub. Also, I employed survival analysis in order to determine the life expectancy of the credit/debit cards produced by the bank.

 
 
 
 
 

BI Analyst Junior

BAC CREDOMATIC

Dec 2014 – Jul 2017 Costa Rica
Credit/debit card fraud detection: I served as a fraud analyst for creating and deploying rules for detecting and declining fraud transactions and avoiding economic losses for the customers. The used techniques for recognizing the anomalies span the areas of supervised and unsupervised Machine Learning.
 
 
 
 
 

Actuarial Intern

National Bank of Costa Rica

Jan 2014 – Jun 2014 Costa Rica
I implemented a software application using Matlab that compares the financial institutes of Costa Rica, using a risk scorecard based on Moody’s methodology.

Accomplish­ments

Data Engineer with Python Track

DataCamp May 2020 – May 2020
See certificate

Expert Program in Data Mining

Promidat Jun 2016 – May 2017
See certificate

Recent Posts

Skill Set for data scientists

On the Internet, you can find a lot of definitions of data science. My preferred description of this science I founded it in the Harvard Data Science Review (HDSR). Rafael A.
Mar 12, 2020 2 min read

Projects

*
All Machine Learning Anomaly Detection UC3M Other

Master’s thesis: Application of Convolutional networks in the Multiple Hypothesis Testing

Site with the official pdf document and Python code of my thesis project for the master’s degree in Statistics for Data Science. My thesis applies Convolutional Neural Networks (CNNs), a popular Deep Learning framework used for detecting patterns, especially in images and photos. But, instead of analyzing pictures, I am applying the CNNs in the context of statistical inference on high-dimensional data and Multiple Hypothesis testing.

Survival Analysis: Hard Drive Reliability Sample

Nonparametric methods such as Kaplan-Meier and Nelson-Aalen is applied to the Hard Drive Data of Backblaze.

Anomaly Detection in Time Series using R

Identifying and predicting anomalies in time series is crucial for decision making. So, we are going to use an option in R for doing the work.

Unsupervised Algorithms

In this project, we review two methods of unsupervised techniques: k-means and hierarchical clustering.

Supervised Algorithms

An introduction to supervised machine learning algorithms, especially methods for classification problems.

Skills

R

SQL

git

Statistics

Machine Learning

Actuarial Science

Contact