Giovanni Spadaro

Areas of study: Machine Learning, Deep Learning, Computer Vision, Natural Language Processing, Statistics, Data Analytics.

Thesis: Predictive algorithms' application in medical diagnosis

Advisor: Prof. Maurizio Palesi

Final grade: 110/110. Areas of study: Computer Science, Machine Learning, Signal Processing, Electronics, Automation.

Supervisors: Prof. Francesco Beritelli, Ing. Michele Russo

Study and implementation of SOTA GAN based architectures for anomaly detection in industrial multivariate time-series datasets. Training and evaluation of such models in order to provide a baseline aimed at predictive maintenance and early machineries fault detection. Technologies: GANs, LSTMs, PyTorch, Scikit-learn.

Designed and implemented a multiclass classification system to evaluate candidate profiles, leveraging advanced feature engineering across both structured datasets and unstructured text sources. Built an interactive Gradio web application to demonstrate the system’s prototype and support rapid model validation and overall feedbacks. Deployed a scalable, production-grade inference infrastructure on AWS, ensuring reliable model serving and smooth integration with existing business workflows. Technologies and Services: AWS (mainly SageMaker, Lambda), Google Cloud, PostgreSQL, Pandas, Scikit-learn, XGBoost, OpenAI APIs, HuggingFace Spaces, Git.

Designing and implementing an autonomous AI agent system to automate businesses workflows, leveraging semantic retrieval and custom Graph RAG architectures to dynamically extract knowledge and construct ad-hoc graphs from unstructured data. Building a production-ready, full-stack interactive web application to serve as the client’s dashboard, enabling real-time agent monitoring, system configuration and workflow validation. Deploying a secure, scalable cloud infrastructure on Azure, utilizing a hybrid relational and graph-based data layer. Technologies and Services: Azure, PostgreSQL, Dgraph, Python, TypeScript, LangChain, Transformers, Pandas, Next.js, Git.

Private lessons of Mathematics, Physics, Computer Science, Electrotechnics, Machine Learning, Natural Language Processing and Computer Vision for MSc, BSc and high school students. Make students grow academically and get them excited about scientific subjects.

Delivered corporate training on Machine Learning, Deep Learning, NLP, Transformers, and Generative AI applications. Covered LLM deployment, RAG architectures, prompt engineering and parameter-efficient finetuning techniques. Included practical sessions on GPU acceleration, finetuning and production AI workflows. Enhanced technical training and knowledge-transfer skills for corporate teams.

Manage Low‑Voltage division software workflow and all team IT. Lead R&D and implementation of the car telemetry platform: embedded data acquisition, CAN integration, telemetry transport, storage, and observability. Define data schemas, coding standards, and documentation; mentor team members. Technologies: C, Git, ESP32, IEEE 802.3cg-2019, CAN bus, Grafana, OpenTelemetry, Next.js, JavaScript.

Series of seminars about Basics of Quantum Mechanics, Quantum Computing and Qiskit framework. Solving 3 sets of challenges using Qiskit.

Understand the basics of cloud computing and modern data architectures and learn about the main AWS technologies for creating and managing Data Storage, Data Processing and Data Analysis steps.

Match analysis, opposition scouting e talent scouting from Volleyball scout datasets.

Series of lessons about Machine Learning, Deep Learning, IoT and Cyber Security from UNICT professors and Samsung experts. Development of an innovative group project: Hexagarden, an intelligent home wall garden.

Introduction to the AWS platform.

Solve a business problem related to the supermarket checkout queue forecast in 32 hours using Machine Learning techniques.