Enterprise Experience

Data Scientist

11-2025/03-2026

Tul SAS

  • Data Warehousing: Designed multiple tables, views, and persistent data structures in Snowflake to support BI consumption, API integration, and ETL pipelines, optimizing query performance and overall efficiency.
  • Report Automation: Built multiple automated reporting workflows using Deepnote and Python, significantly reducing report generation time and enabling data-driven business decisions with minimal manual intervention.
  • RAG (Retrieval-Augmented Generation): Implemented a RAG system for image recognition using a Large Language Model (LLM) to extract key billing information from invoices, eliminating full manual transcription and improving process efficiency.
  • DBT Projects: Designed and executed DBT + Snowflake projects to efficiently transform and model data, enhancing data quality and accessibility for analytics and decision-making.
  • Microservices Data Modeling: Designed efficient data models for microservices architectures using SQL and Snowflake, ensuring application scalability and performance.
  • Agile Methodologies: Worked in an Agile development environment using Jira, collaborating with cross-functional teams to deliver data projects iteratively and efficiently.
  • Version Control with Git: Leveraged Git and Bitbucket for source code version control, streamlining collaboration and change tracking across data projects.
  • BI Data Provisioning: Prepared and structured data for consumption by Business Intelligence tools, ensuring clean, well-organized, and readily accessible datasets for analysis and visualization.

Senior Reporting Analyst

03-2024/11-2025

Emergia Contact Center

  • Communication Skills: Successfully collaborated with four operations managers from Seguros Bolívar, Credivalores, Jelpit, and Bancolombia, demonstrating strong stakeholder communication and cross-functional coordination abilities.
  • Compelling Visuals: Designed and automated an aesthetically polished Power BI dashboard covering over one million monthly transactions, leveraging SQL Server as the data backbone.
  • Process Automation: Streamlined Credivalores' Emmaster weekly hours calculation and extraction process from 2 hours down to just 10 minutes using SQL and .NET.
  • ETL Pipeline Development: Architected 10+ multi-step ETL pipelines using SQL and Apache Spark (PySpark) to effectively manage and control end-user data flows.
  • Data Engineering: Automated Bancolombia's database loading workflows using Python and data pipelines for segmentation and processing, reducing manual effort from 11 hours to just 1.5 hours.
  • NoSQL Handling: Consistently performed accurate JSON-to-SQL conversions to process and interpret click-to-call lead data from client sources.
  • Business Intelligence: Designed 10+ Power BI dashboards and 20+ paginated reports featuring key business KPIs to support effective marketing and corporate decision-making.

Geothermal Plant Designer

08-2023/02-2024

CHEC Grupo EPM

  • Utilized pivot tables and Tableau to summarize critical inventory and cost information.
  • Effectively compared various project alternatives.
  • Designed distribution and transmission networks for connecting the geothermal power plant to the national grid (SIN).
  • Conducted technical reviews of electrical connection elements, including single-line diagrams. Performed calculations related to regulation, losses, and inventory.
  • Developed a complete financial model for the project. Evaluated project viability using metrics such as NPV, IRR, and LCOE.
  • Calculated the weighted average cost of capital (WACC) effectively. Considered five distinct sources of capital expenditures (CAPEX).
  • Georeferenced structures and optimized the existing medium-voltage network.
  • Utilized AutoCAD, Google Earth, and PLS-CADD to efficiently locate support structures. Ensured compliance with safety distances specified by RETIE (Colombian electrical code).
  • Employed descriptive and machine learning models, including the powerful LightGBM model. Engineered relevant features for accurate price predictions in the energy market.

Freelancer Experience

Asset Manager

04-2020/Present

Nexoral Group LLC

  • See my performance at THIS LINK
  • Investment Performance: Achieved outstanding investment results, consistently outperforming the S&P500 with an annual effective return of 28.22% over the past four years. This performance underscores the stability and rapid growth potential of my portfolio.
  • Risk-Reward Optimization: Demonstrated proficiency in optimizing risk-reward trade-offs by combining Markowitz’s efficient frontier theory, Python programming, and cross-validation techniques. These efforts ensure the optimal allocation of asset weights.
  • Advanced Python Skills: Leveraged Python for finance to consume financial data via APIs and fine-tune asset weights for optimal returns.
  • Kaggle Competition Success: Awarded the silver medal in the Kaggle competition “Optiver Trading at the Close.” Employed powerful time series forecasting models, feature engineering pipelines, and tools such as Pandas, Polars, LightGBM, and Optuna to achieve precise results.
  • Financial Accounting Expertise: Proficiently interpreted key metrics from balance sheets, income statements, and cash flow reports. Familiar with indicators including ROE, ROA, current ratio, and price ratios (e.g., PER, EV/EBITDA).
  • Corporate Governance and Financial Mathematics: In-depth knowledge of corporate governance practices, present and future values, depreciation, contingencies, and financial mathematics.
  • Developed a winning stock options strategy by selecting those with the best risk-reward ratio. Created a Python program to automate the process of identifying the optimal options to buy.

Machine Learning Contributor

09-2023/01-2024

Kaggle

  • Attained a silver medal in Optiver Trading at the Close competition, showcasing proficiency in algorithmic trading strategies.
  • Conducted forecasting of 200 stock prices using a vast dataset, demonstrating expertise in financial data analysis and prediction.
  • Designed and implemented a comprehensive feature engineering, training, and inference pipeline on Kaggle, optimizing model performance and accuracy.
  • Conducted fine-tuning of LightGBM, a leading model for tabular data, to achieve superior predictive performance in various domains.
  • Trained LLM models, including the Deberta V3 transformer, for advanced natural language processing tasks, exhibiting proficiency in cutting-edge deep learning techniques.
  • Demonstrated strong proficiency in feature engineering principles and the determination of feature importances, enhancing model interpretability and performance.
  • Possess a robust understanding of time series forecasting methodologies, enabling accurate prediction and analysis of sequential data patterns.
  • Proficiency use of TensorFlow, Keras, Sklearn and pyTorch APIs.
  • Developed AI for image labeling and segmentation using transformers for computer vision.