Projects
Here you'll find a collection of my recent projects that demonstrate my expertise in AI engineering, machine learning, and sustainable technology solutions. Each project reflects my commitment to creating impactful, technically sophisticated applications.
Text Summarizer
10.2024This RAG system summarizes any text of any size by using real-time indexing.
Chat with your data
09.2024This self-corrective RAG system can take an upload of any text and answer questions about it.
Day-Ahead PV power production forecast
02.2024This RNN model is able to predict the PV power production for tomorrow (24 hours) based on the existing information available up until 12:00 of the previous day.
Featured Project Details
Text Summarizer
This advanced RAG (Retrieval-Augmented Generation) system efficiently summarizes text of any length using real-time indexing techniques.
Technical Highlights:
- Implemented a custom chunking strategy to handle documents of varying sizes
- Built a real-time indexing pipeline using ChromaDB for efficient vector storage
- Optimized prompt engineering to generate coherent, comprehensive summaries
- Integrated evaluation metrics to assess summary quality and relevance
Chat with your data
A self-corrective RAG system that enables users to upload any text document and interactively ask questions about its content.
Technical Highlights:
- Developed a self-correction mechanism to improve answer accuracy
- Integrated multiple LLM models for different reasoning steps
- Implemented document preprocessing pipeline for optimal information extraction
- Designed an intuitive chat interface for seamless user interaction
Day-Ahead PV power production forecast
A recurrent neural network model that accurately predicts photovoltaic power production for the upcoming 24 hours based on data available up until 12:00 of the previous day.
Technical Highlights:
- Designed and trained an RNN model optimized for time-series forecasting
- Incorporated weather prediction data to improve forecast accuracy
- Implemented feature engineering to capture solar irradiance patterns
- Developed a Streamlit web application for real-time forecasts visualization
Interested in collaboration?
Let's discuss how we can work together on innovative AI solutions.