Hi there! 👋 I'm Cristian, a recently graduated Systems Engineer passionate about building software solutions that combine backend development, data engineering, and machine learning.
My main interest is developing end-to-end systems that transform data into valuable products. I enjoy working across the entire software lifecycle: designing APIs, building web applications, developing data pipelines, training machine learning models, implementing automated testing, and deploying solutions in reproducible environments.
Throughout my academic and personal projects, I have developed applications ranging from healthcare decision-support systems powered by machine learning to data platforms built with modern ELT architectures and AI-powered educational applications.
💡 What motivates me: solving real-world problems through software. I am particularly interested in backend development, data-intensive applications, and intelligent systems that create measurable impact.
🌍 Portfolio: https://portafolio-eight-inky-71.vercel.app
📊 Kaggle: https://www.kaggle.com/criser2013
Python • SQL • JavaScript • Java • R
FastAPI • Flask • Django REST Framework • Express.js
React • Vite • HTML • CSS
PySpark • Apache Airflow • dbt • Pandas
Scikit-learn • PyTorch • MLlib • SHAP • LIME
PostgreSQL • MySQL • Firestore
Docker • Git • GitHub Actions • Pytest • Jest • Playwright
Power BI • Microsoft Excel
Google Cloud Platform (GCP) • Microsoft Azure
An end-to-end intelligent healthcare platform designed to support pulmonary embolism (PE) diagnosis through machine learning and explainable AI.
- Developed and validated machine learning models using clinical data from 161 patients and 44 diagnostic features.
- Evaluated 2.7+ million model configurations across multiple algorithms including XGBoost, Random Forests, Neural Networks, and other ML techniques.
- Achieved AUC = 0.93 and F1-score = 0.92 during internal validation.
- Performed external validation using an independent dataset of 128 patients.
- Applied explainability techniques with SHAP and LIME to interpret model behavior and individual predictions.
- Exported the final model to ONNX for optimized inference in production.
- Developed a full-stack web application for patient management, diagnosis visualization, and result export.
- Built a FastAPI backend integrating authentication, machine learning inference, and patient data management.
- Integrated Google Drive API and Firestore for secure data storage and collaboration.
- Implemented automated testing with Pytest and Jest.
- Containerized application services with Docker.
Tech Stack:
Python, FastAPI, Scikit-learn, ONNX, SHAP, LIME, React, Vite, Firebase, Firestore, Google Drive API, Docker, Pytest, Jest
Repositories:
An end-to-end data engineering project that transforms semi-structured transactional data into a business-ready analytical model.
- Designed and implemented an ELT pipeline following the Medallion Architecture (Bronze, Silver, Gold).
- Ingested JSON data from AWS S3 into a scalable analytical workflow.
- Processed and transformed data using PySpark.
- Applied cleaning, deduplication, missing-value handling, and business transformations.
- Built dimensional models using a Star Schema approach.
- Implemented data quality validations and transformation workflows with dbt.
- Created Power BI dashboards capable of answering 24 business questions from the transformed data.
Tech Stack:
PySpark, Apache Airflow, dbt, PostgreSQL, Docker, AWS S3, Power BI
Repository:
A natural language processing project focused on identifying the best neural architecture for multiclass sentiment classification in Spanish.
- Processed and analyzed over 200,000 Amazon reviews written in Spanish.
- Compared traditional and embedding-based text representations.
- Trained and evaluated MLP, RNN, LSTM, and GRU architectures.
- Fine-tuned Transformer-based models including BETO and DistilBETO.
- Compared performance, training efficiency, and generalization capabilities across architectures.
- Identified DistilBETO as the best-performing model for the task.
Tech Stack:
PyTorch, Transformers, NLTK, Gensim, FastText, Pandas, Power BI
Repository:
A web platform that personalizes English learning activities using generative AI and speech technologies.
- Developed personalized English-learning experiences based on user interests and proficiency levels.
- Integrated generative AI to create educational content dynamically.
- Implemented pronunciation assessment, dictation exercises, guided conversations, and adaptive quizzes.
- Integrated Gemini API and Azure Speech Services for content generation, speech recognition, and pronunciation evaluation.
- Developed authentication and data management features using Firebase services.
Tech Stack:
React, Vite, Firebase Authentication, Firestore, Gemini API, Azure Speech Services
Repository:
A software quality engineering project focused on functional, non-functional, and automated testing.
- Designed and executed functional test suites using black-box testing techniques.
- Developed automated tests using Playwright.
- Implemented unit tests with PHPUnit achieving over 90% code coverage.
- Evaluated usability and accessibility using WCAG 2.2 guidelines and Google Lighthouse.
- Performed load and performance testing with Apache JMeter.
https://portafolio-eight-inky-71.vercel.app/work/dolibarr
Tech Stack:
Playwright, PHPUnit, Apache JMeter, Google Lighthouse
Repository:
A complete ETL and analytics project focused on transforming operational data into actionable business insights.
- Built an end-to-end ETL workflow for the Adventure Works dataset.
- Performed data cleansing, transformation, and normalization.
- Integrated Azure Translation Services for multilingual data processing.
- Designed and published Power BI dashboards for business KPI monitoring and analysis.
Tech Stack:
SQL Server, PostgreSQL, Pandas, Azure Cognitive Services, Power BI
Repository: