My Portfolio
Portfolio
My Latest Projects
Take a look at my recent work.
Natural Language Processing, NLP
Math Problem-Solving Agent
An advanced AI-powered solution designed to tackle complex mathematical problems with precision and efficiency. This agent employs sophisticated reasoning frameworks and iterative refinement techniques to ensure accurate and reliable answers, making it an invaluable tool for both students and professionals in need of quick, dependable problem-solving.
Social Media Content Agent
A dynamic AI-driven solution designed to streamline content creation across multiple platforms. This agent efficiently generates blog posts, LinkedIn updates, and tweets by analyzing and summarizing relevant YouTube videos on your chosen topics, ensuring that your social media presence remains engaging and up-to-date with minimal effort.
TextSummarizer Pro is a robust text processing application designed for efficient document summarization. Leveraging advanced tools like LangChain and OpenAI’s GPT-3.5-turbo, the app utilizes a sophisticated MapReduce framework to process and condense large text documents into concise summaries. This solution is ideal for users needing to distill large volumes of information into digestible insights.
QueryMaster AI is an intuitive app designed to help you quickly find answers to your questions from large amounts of text. It combines smart technologies like LangChain, OpenAI’s powerful language model GPT-3.5-turbo, and Pinecone’s fast search database to deliver accurate results. The app takes your question and looks for the most relevant parts of the text using OpenAI’s language understanding. It then uses a special process to read through the important parts and find the best answer for you.
Advanced Time Series Forecasting Portfolio
Pioneered cutting-edge forecasting solutions across diverse domains, leveraging sophisticated statistical methods and machine learning techniques to deliver unparalleled predictive accuracy:
- Award-Winning Financial Time Series Analysis: Spearheaded groundbreaking research in financial time series, earning recognition as the best Master’s thesis at the Instituto Politecnico Nacional. Demonstrated exceptional analytical skills and innovative approaches to complex economic data.
- Multi-National Consumer Price Index Forecasting: Engineered a robust comparative study of ARIMA models versus Ensemble Machine Learning techniques for predicting Consumer Price Indices in Germany, France, and the United States. Uncovered key insights into international economic trends and forecasting methodologies.
- Large-Scale Retail Demand Forecasting: Architected a scalable Multiple SKU time series forecasting system for a major supermarket chain, significantly enhancing inventory management and reducing stockouts across thousands of products.
- AI-Powered Electric Load Prediction: Developed an advanced electric load forecasting model integrating natural language processed features with state-of-the-art quantile regression convolutional and transformer neural networks. Achieved unprecedented accuracy in predicting energy demand fluctuations.
- Market-Driven Energy Price Forecasting: Revolutionized Local Marginal Price predictions in the Mexican Electricity Wholesale Market by implementing a hybrid model combining Deep Learning and Autoregressive Vectors. Empowered stakeholders with highly accurate price forecasts, facilitating optimal decision-making in a volatile market environment.
Windrose — Canadian Precipitation Network Audit
Windrose: The first independent, station-level reliability assessment of Environment and Climate Change Canada’s precipitation monitoring network. Published April 16, 2026. Covers 8,374 stations across 12 provinces and territories using a methodology specified by the office of Minister Julie Dabrusin.
Key findings: 81.2% of ECCC stations are decommissioned. 1 in 4 active stations has confirmed transmission issues. 101 ghost stations identified. Zero EXCELLENT-rated stations in NWT or Nunavut. Average data quality: 52.95%.
Applications: Parametric insurance triggers, crop insurance actuarial models, hydrology and flood forecasting, engineering design standards, and federal climate reporting.
Learn more: windrose.news