Marcelo Botero Gómez Email & Phone Number

Bogota, D.C., Capital District, Colombia

I lead Siigo’s Data Science front. I structured the Data Science area from its creation, and have led the design, development, and deployment of over twenty high impact projects, including:- ML and Statistical Analysis projects: Clustering models for Client Profiling and Database Optimization, enabling streamlined customer provisioning. Predictive algorithms for Churn and Payment Forecasting, ensuring effective customer retention strategies. Technology Macro and Micro costing models, optimizing expense management and enabling strategic decision-making.- AI Projects: Customer Interactions AI-Driven Insights - AI-driven pipeline for vast customer interaction data, extracting insights using speech-to-text and text classification technologies. Customer profile AI-powered building - On demand pipeline using NLP to extract client transactions from software interactions and AI powered analysis to build customer profiles from most frequent transactions.- Optimization Projects: Workforce Management: Network flow optimization model automating schedule assignments, achieving remarkable run times for companywide schedule generation.

Bogota,D.C., Capital District, Colombia

As the company’s first Data Scientist, I brought advanced analytics into the Data Team’s landscape. In the six months within this role, I developed two high impact projects:- Customer Segmentation Pioneering: Structuring and cluster analysis of usage and customer interaction data, unveiling insights into customer behavior and correlation with user activity.- Cross-sell propensity model: Classifier model to identify profiles of existing payroll product buyers and detect similar patterns in potential cross-sell customers boosting cold-call strategies.

Bogota,D.C., Capital District, Colombia

Helped drive business value through feature engineering, data visualization and analysis. Developed and deployed high impact Power BI dashboards for M&A, sales and operations teams.

Colombia

Brought advanced data analytics to the zonal division of Bogotá’s SITP, designing, developing and deploying two main projects:- Calculation of travel demand through OD matrix construction, using big data technologies to process and model the ticket registry of the city, with more than 3+ million daily transactions. Developed a methodology to estimate travel demand in each bus route, using probabilistic models and inference techniques. Optimized algorithm logic to enable the model to run within a common desktop computer, with excellent sub-hour run times.- Optimization network flow model for schedule automation of control center employees, using demand curves and operational data to generate model parameters. Designed the model with Pyomo, taking advantage of open solvers to deploy a service-ready model for operation.

Bogota,D.C., Capital District, Colombia

Assisted the transformation to data-driven decision making of the company, co-creating the central analytics platform and deploying three high impact projects:- Anomaly Detection algorithm for identification of transactions under AML/CFT regulations, utilizing machine learning techniques to derive best results.- Construction of Risk mitigation models using network science to boost market operations.- Unification of descriptive, predictive and prescriptive models within a single platform deployed with Shiny R, enabling the whole company from operational to executive levels to switch to data-driven real time decisions.

Colombia

Managed the administrative, accounting and legal branches of a food industry startup, as well as the operations department, boosting efficiencies using data analytics. Implemented geographical analysis to boost market strategies and conceptualized an integrated analytics portal for allies.

Bogotá, Bogotá D.C. Department, Colombia

Assisted in COPA and EpiAndes investigation groups. Co-authored an academic paper in transport analytics, through the reviewing of academic literature and development of clustering algorithms for data-driven-LTS. Built tools to predict road categories in low computational time.