Portugal
Hi there! I'm a graduate MSc student in Computer Science and Engineering @ NOVA SST, with a deep passion for both technology and finance. My academic journey has been driven by curiosity, and I've had the chance to dive deep into Explainable AI, Causality and Graph Neural Networks during my research projects and my Master's thesis. I am passionate about data science, with a focus on deep learning and finance. While I have a particular interest in using AI for algorithmic trading and quantitative modeling, I enjoy working across the entire data science pipeline. This includes collecting, processing, building, training, and deploying models in various domains to solve complex problems and generate actionable insights. In addition to my interest in AI and finance, I also enjoy building software across various platforms, whether it’s web, desktop, or mobile. I’m always excited to learn new tools and frameworks, and I thrive on the challenge of creating efficient and innovative solutions. When I’m not working, you can find me playing football or learning new finance concepts! I'm always eager to learn and collaborate on innovative projects at the intersection of AI, data science, and finance, as well as web, desktop, and mobile development. Feel free to connect if you're interested in exploring opportunities or exchanging insights in these fields!
- Incorporating in Data Engineering, Machine Learning, and Data Science teams, gaining hands-on experience with scalable data pipelines, real-world ML models, and advanced analytics for business decision-making. - Improved the reliability of concert-revenue forecasting in the Data Science team by building a data and model drift monitoring system to detect distribution shifts and support city-level marketing budget recommendations. - Building a production-grade training pipeline in the ML Engineering team, with Airflow, MLflow, and versioned datasets, automating feature engineering, model training, validation, and deployment for city-level marketing budget optimization.
- Built, published, and deployed a reusable React component library (TypeScript+Vite) across 3 services, improving UI consistency, development speed, and maintainability via full CI/CD pipeline while working in Agile. - Enhanced a client retention service endpoint using Java, Spring AI, and Ollama to generate AI-powered, user-friendly summaries of configuration validation errors and next actions to take to solve those errors. - Added client retention service endpoint with Java, Spring AI, and Ollama, generating AI-powered retention configuration suggestions to help users make faster, informed decisions, enhancing engagement.
- Worked on a GNN approach to space traffic management, specifically, collision avoidance. - Collected publicly available TLE data to create a large time-series dataset. - Built a discrete-time dynamic space conjunction graph based on a transformed evenly spaced over time TLE dataset. - Achieved 80% test accuracy in conjunction forecasting using RNN/CNN-based Spatio-Temporal GNN, outperforming a heuristic model by 10x.
- Researched various methods and papers in Human-Centered Explainable AI. - Built a pipeline to convert explanation subgraphs generated by XAI methods into human-readable text using LLMs. - Designed a user study to evaluate the quality of generated explanations for referral recommendations from GNNs.
- Processed a large dataset from a private healthcare provider to build a doctor referral graph for a Graph Neural Network. - Trained a Graph Neural Network for doctor referral recommendation, achieving 81% test accuracy. - Developed and implemented a novel causal explanation method for doctor referral recommendations. - Implemented two baselines for comparison and demonstrated a 4x performance improvement over baseline methods. - Wrote a research paper about the novel causal explanation method, supporting XAI-driven decision-making in healthcare.