Hi there! 👋 I'm Cristian, a Systems Engineer bridging the gap between Data Science and Full-Stack Web Development.

I specialize in building intelligent applications—designing robust backend systems with Python, crafting intuitive user interfaces with React, and deploying machine learning models into production environments using Docker and Cloud services (GCP & Azure).

💡 What drives me: I love turning complex datasets into actionable insights and real-time interactive tools. I'm a self-motivated learner, constantly exploring new technologies to solve real-world problems.

🏆 Certifications: Certified IBM SkillsBuild Data Analytics & Google Cloud Computing Foundations.

🌍 Check out my Portfolio | 📊 Find me on Kaggle

• Programming languages: Python, JavaScript, Java, R, SQL
• Backend: FastAPI, Django REST, Express.js, Flask
• Frontend: React + Vite, HTML, CSS
• Data: Pandas, PySpark, dbt, Apache Airflow, Matplotlib, Seaborn
• AI: Scikit-learn, PyTorch, MLlib
• Databases: PostgreSQL, MySQL, Firestore
• DevOps: Docker, Git, GitHub Actions
• Testing: Pytest, Playwright, Jest, Selenium IDE
• BI: Power BI, Microsoft Excel
• Cloud providers: Google Cloud, Microsoft Azure

• Developed a machine learning system to assist in PE diagnosis.
• Executed the entire data science pipeline: exploratory data analysis (EDA) on source data, cleaning and preprocessing training/validation data, and model tuning.
• Assessed model performance using cross-validation and an external dataset to verify model generalization capabilities.
• Analyzed model outputs using SHAP explainers alongside a lightweight LIME approach optimized for the production environment.
• Developed a web application to manage diagnoses and interact with the machine learning model in real-time.
  • Stack: React, FastAPI, ONNX, Firebase, Docker, Scikit-learn, GitHub Actions, Google Drive API, reCAPTCHA, SHAP, LIME, Pytest, Jest
  • Repositories: Frontend, Backend, and Model Development

• Trained and evaluated four neural network architectures (MLP, RNN, LSTM, GRU) and Transformers for a multiclass classification task.
• Processed and analyzed over 1.2 million Amazon reviews using NLP techniques and different word representations (TF-IDF and embeddings).
• Compared the performance of TF-IDF, self-trained embeddings, and embeddings pre-trained on the Spanish Billion Corpus.
• Conducted a comprehensive performance assessment to select the most accurate architecture for review classification.
  • Stack: PyTorch, Pandas, Transformers, NLTK, FastText, gensim
  • Repositories: Models

• Developed a web application for creating and customizing English learning activities tailored to the user’s proficiency level and preferred topics.
• Implemented interactive features such as dictation, pronunciation practice, guided conversations, and dynamic quizzes.
• Integrated AI services for automated content generation and evaluation, including pronunciation analysis and educational content creation.
• Implemented user authentication with Firebase Auth, data management with Firestore, and integrations with the Gemini and Microsoft Azure Speech APIs.
  • Stack: React, Firebase, Microsoft Azure Cognitive Services, Gemini API
  • Repositories: Application

• Designed and executed functional tests for invoicing, ticketing, and MRP modules, applying black-box techniques such as equivalence partitioning and decision tables using Playwright (Python).
• Developed unit tests with PHPUnit for critical system classes, applying statement, branch, and condition coverage criteria to achieve over 90% code coverage.
• Evaluated non-functional quality attributes, focusing on accessibility and usability via WCAG 2.2-based heuristics validated with Google Lighthouse.
• Performed load and performance testing using Apache JMeter within a Dockerized environment.
  • Stack: Playwright, PHPUnit, Apache JMeter, Google Lighthouse
  • Repositories: Tests

• Built an end-to-end ETL pipeline for the "Adventure Works" database, applying data cleaning, structural normalization, and transformation logic.
• Integrated Azure Cognitive Services (Translation API) for multilingual data processing.
• Designed and published interactive reports in Power BI to track and visualize strategic business KPIs.
  • Stack: SQL Server, PostgreSQL, Pandas, Microsoft Azure Cognitive Services, Power BI
  • Repositories: ETL

• Conducted EDA on the Vinho Verde dataset, verifying variable distributions, identifying outliers, and analyzing attribute correlations.
• Trained and evaluated multiclass classification models using algorithms like KNN, SVM, Random Forests, and Neural Networks.
• Applied feature engineering methods and oversampling techniques to resolve class imbalance and optimize classifier performance.
  • Stack: Python, Pandas, Scikit-learn, Matplotlib
  • Repositories: Models