Juan Jose Restrepo Rosero
Software Engineer, Data Scientist
- PHONE:
- +57 316-617-3202
- EMAIL:
- restrepojuanjo@gmail.com
- LOCATION:
- Cali, Colombia
About
Trilingual Electronics Engineer specialized in backend development, data engineering, and machine learning, currently pursuing a Master’s degree in Data Science. Experienced in building production-ready systems, including RESTful APIs, data pipelines, and predictive models. I apply modern software engineering practices such as static code analysis, type safety, and modular design to deliver scalable, maintainable, and reliable solutions. My work focuses on end-to-end systems, from data processing and modeling to backend integration and deployment. I enjoy working on multicultural teams and always stay motivated to learn new things and approach challenges with a positive and creative thinking.
Download CVServices
Software Development
Full-stack development with modern technologies. Expertise in Python, JavaScript, and various frameworks for building scalable applications.
Data Science & ML
Data analysis, machine learning model development, and implementation of AI solutions using R, Python, TensorFlow, and related technologies.
DevOps Engineering
Implementation of CI/CD pipelines, infrastructure automation, and cloud solutions using modern DevOps practices and tools.
Robotics & Automation
Design and implementation of robotics systems and automation solutions, leveraging mechatronics and electronics engineering background.
Resume
EDUCATION
PONTIFICIA UNIVERSIDAD JAVERIANA CALI
July 2024 - Present
Master’s Degree in Data Science
PONTIFICIA UNIVERSIDAD JAVERIANA CALI
January 2019 - April 2024
Bachelor of Science in Electronics Engineering
SOUTHERN LEE HIGH SCHOOL
August 2017 - June 2018
Foreign Exchange Student
SAN ANTONIO MARÍA CLARET
September 2011 - June 2017
High School Student
WORK EXPERIENCE
VEEVART
December 2024 - May 2025
Software Engineer I
HOLCIM ABS
July 2023 - January 2024
Engineering Intern
PONTIFICIA UNIVERSIDAD JAVERIANA CALI
August 2022 - December 2022
Laboratory Practices Monitor/Assistant
Skills
DATA ENGINEERING & ML
PostgreSQL
dbt
Scikit-learn
TensorFlow
BACKEND DEVELOPMENT
TypeScript
REST APIs
LANGUAGES
SQL
C++
R
TOOLS & CLOUD
Power BI
PORTFOLIO
WORLD CUP PREDICTION ENGINE (END-TO-END MLOPS)
Production-grade MLOps ecosystem featuring a Segment-Aware Hybrid Ensemble with specialized Draw-Specialist routing.
It implements a monthly Automated Retraining Pipeline via GitHub Actions, integrating real-time data from Supabase hosted PostgreSQL.
Features a Champion vs Challenger promotion gate with automated regression blocking, and Shadow Deployment for real-time experimental model validation.
The architecture follows a strict Medallion Data Design (Bronze/Silver/Gold) powered by dbt for data quality, achieved 100% MyPy Strict type-safety, and includes a Dynamic Theme Engine for high-contrast accessibility.
USED STACK:
F1 2026 SEASON PREDICTIVE PLATFORM (END-TO-END MLOPS)
A production-grade, end-to-end MLOps platform designed to predict Formula 1 race dynamics for the 2026 regulation era. This system combines state-of-the-art Gradient Boosting (XGBoost/LightGBM) with a high-fidelity interactive dashboard inspired by F1 TV telemetry.
Features Autonomous Orchestration via Trigger.dev for Friday forecasts and Monday audits, and AI-Generated Race Narratives powered by Gemini 2.5 Flash that synthesize telemetry residuals into professional engineering briefings.
USED STACK:
FOOTBALL ANALYSIS & PREDICTOR (EURO, COPA AMÉRICA, QATAR 2022)
Comprehensive data science project focused on analyzing team performances and predicting outcomes for major tournaments like UEFA Euro 2024, Copa América 2024, and FIFA World Cup Qatar 2022. Implements end-to-end workflows including automated data collection, cleaning, exploratory analysis (EDA), and statistical modeling to forecast results and highlight trends.
USED STACK:
TASK MANAGER API (PRODUCTION BACKEND)
Production-ready RESTful API (CRUD) built with Node.js, Express,
TypeScript, Prisma, and PostgreSQL. Implements JWT-based
authentication, rate limiting, schema validation (AJV), and
centralized error handling following robust backend engineering
practices. Fully documented with Swagger and deployed on Render.
Live Status: Production (Render) – Available
USED STACK:
Knowledge Management & Multi-Spectroscopy Data Pipeline
End-to-end ELT pipeline for multi-technique spectroscopy data (Raman, FTIR, UV-Vis), developed using Python, Airflow, PostgreSQL, and dbt. Designed to transform fragmented and heterogeneous research data into structured, reliable, and actionable datasets. Implements Medallion Architecture, data validation, and reproducible workflows to support scientific analysis and decision-making.
USED STACK:
Qversity Data Engineering Pipeline
Production-grade ELT pipeline implementing Medallion Architecture, using Airflow, dbt, and PostgreSQL. Integrated data ingestion from AWS S3 and containerized deployment with Docker. Ensured data quality through testing, validation, and structured transformations, delivering actionable business insights such as revenue segmentation, ARPU analysis, and customer behavior patterns.
USED STACK:
DENGUE ICU PEDIATRIC ANALYSIS (CLINICAL DATA SCIENCE)
End-to-end data science project analyzing pediatric dengue cases in ICU,
Focused on identifying clinical risk factors associated with mortality and
severe outcomes in pediatric ICU dengue cases (~200 patients). Combines
rigorous EDA and inferential statistics (Mann-Whitney U, Chi², Spearman)
with an interpretable XGBoost model using SHAP.
Built with production-grade practices (modular OOP, Pandera validation,
reproducible pipelines). Key predictors include BUN, AST, vasopressor
usage, and thrombocytopenia, consistent with dengue severity patterns.
USED STACK:
SENTIMENT ANALYSIS
Natural Language Processing (NLP) pipeline for sentiment classification on social media data using classical machine learning and deep learning models. Includes data preprocessing, feature extraction, and model evaluation, with applications in healthcare-related analytics and public perception monitoring.
USED STACK:
Repository
AUTOMATED MANUFACTURING CELL CONTROL SYSTEM
Developed a control system for an automated manufacturing cell, integrating robotics, PLCs, and HMIs to optimize industrial processes. Focused on real-time control, system integration, and industrial automation reliability in a production-like environment.
USED STACK:
AUTONOMOUS NAVIGATION ROBOT
Designed and implemented an autonomous navigation system using computer vision and machine learning techniques. The system performs environment mapping, obstacle avoidance, and path planning through sensor fusion and SLAM-based approaches.
USED STACK:
UNO CARD GAME (PYTHON)
Object-oriented implementation of the UNO card game with complete game logic and CLI interaction.
USED STACK:
RepositoryGet in touch_
- PHONE:
- +57 316-617-3202
- EMAIL:
- restrepojuanjo@gmail.com
Or just send me an email here