Hey there, I'm Daniele! I'm currently diving deep into the world of Artificial Intelligence as a Master's degree student. I love learning about new software technologies and apply them to solve problems and create new projects. I'm currently interested in NLP, Deep Learning and anything related to AI. In my free time, I like to engage in sports 🏃, read books 📖 and play the guitar 🎸.
- Design, implementation, and maintenance of Artificial Intelligence applications for internal and external clients. Integration of AI functionalities into existing software by collaborating with the respective corporate development teams.
- Preparation, management, and manipulation of data necessary for training and running AI systems in collaboration with the areas/clients that produce them.
- Development of my thesis project, which aimed to process custom data of various types (textual and tabular data) to make them usable by LLMs.
- Throughout this period, I used the OpenAI API and extensively tested numerous open-source language models.
- Conducted fine-tuning experiments and compared their effectiveness with a retrieval augmented generation approach.
It's a two-year Master in Artificial Intelligence which provides solid competence and expertise in the founding areas and innovative applications of Artificial Intelligence. Graduated with a final score of 110/110 with honors.
I made a lot of projects during the degree course and I learnt to work in a team to achieve a common goal and by comparing my ideas with others. Graduated with a final score of 108/110.
In the last 2 years of High School I became passionate about Computer Science and I decide to go on studying this topic.
I first performed an analysis of the lyrics using LSA and other techniques in R. Then I implemented different GNNs, on a narrow subset of songs, that are able to recommend the most similar songs to a given one.
We implemented different pipelines in order to retrieve the most relevant documents, from a subset of ClueWeb12, given some comparative questions. In another task we performed stance classification on these documents.
In one assignment we implemented different models for POS tagging, in the other one some models for abstractive QA on CoQA.
We implemented 3 different models for the 'Combinatorial and Decision Making Optimization' exam using CP, SMT and MIP techniques to solve the VLSI design problem.
I implemented a deep learning model, for the 'Deep Learning' exam, that allows to separate 2 mixed images taken from MNIST and FASHION MNIST datasets.
The system has been developed as my graduation project. The aim of the project is to fight and avoid clothing's counterfeit thanks to the use of blockchain technology.
This is a generic scoreboard that you can use for any game you are playing with your friends. You can choose how many points to reach and whether who has more or who has less wins
A simple script to check if messages containing certain words arrive on your telegram account. A song is played to alert you or to wake you up in case you are waiting for an important message during the night.