Skip to content
View lucalullo's full-sized avatar
525 followers · 2k following

Block or report lucalullo

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
lucalullo/README.md

Luca Lullo

Data Science & AI

I build open-source and personal projects in data science, machine learning, and AI, ranging from applied analysis to experimental libraries and implementations of machine learning systems from first principles.

I mainly work with Python and SQL, focusing on data preparation, feature engineering, model evaluation, and reproducible experimentation. I also explore language models, AI agents, and experimental approaches to tabular machine learning.

Across my projects, I emphasize testing, documentation, reproducibility, and clear communication of methods, results, and limitations.

Kaggle

Expert in Datasets and Notebooks

I publish public datasets and notebooks focused on data quality, exploratory analysis, and predictive modeling.

Tools and libraries

Tools used across my projects and coursework:

Languages: Python, SQL
Data analysis: pandas, NumPy, Jupyter
Machine learning: scikit-learn, XGBoost, LightGBM, SHAP
Deep learning: PyTorch, TensorFlow
Version control: Git, GitHub

Selected projects

My repositories include experimental libraries, educational implementations, and applied data analysis projects.

Project Description
CodAdapt Experimental library for tabular binary classification and regression, based on adaptive coded memory.
NeuroTabular Experimental PyTorch library for tabular binary classification, with preprocessing and a scikit-learn-style API.
BurOCRazia Local application for identifying Italian administrative documents, using OCR and a traceable SQLite catalog.
Building Agentic AutoML Experimental project exploring agent-driven machine learning workflows and iterative evaluation.
Building AI Agent Educational Python project covering agent routing, planning, tool execution, and memory.
Building LLM Educational progression from statistical language modeling to a small decoder-only Transformer.
Building Decision Trees Educational implementation of decision trees from scratch, progressing from simple stumps to recursive trees, missing-value handling, and gradient boosting.
Customer Support Agent Agent project combining request classification, automated handling, and human escalation.
Home Credit Default Risk Credit-risk modeling project with feature engineering, gradient boosting, and SHAP-based interpretation.
Global Emissions & Temperature (1950-2024) Analysis of public data on CO2 emissions, greenhouse gases, and global temperature trends.

Certificates

I complement my projects with structured courses and certificate programs:

  • Google Data Analytics Professional Certificate, Google
  • Production Machine Learning Systems, Google Cloud
  • Supervised Machine Learning: Classification & Regression, IBM
  • Getting Started with TensorFlow 2 and Customising Your Models with TensorFlow 2, Imperial College London
  • Natural Language Processing on Google Cloud, Google Cloud
  • Computer Vision Fundamentals with Google Cloud, Google Cloud
  • Databases and SQL for Data Science with Python, IBM

Connect

LinkedIn | Kaggle

Pinned Loading

  1. codadapt codadapt Public

    Experimental tabular machine-learning algorithm based on adaptive coded memory and shared multi-resolution encoding.

    Python

  2. neuro-tabular neuro-tabular Public

    Experimental neural network library for tabular data with a scikit-learn-style API and automatic preprocessing.

    Python

  3. burocrazia burocrazia Public

    Applicazione locale e deterministica per riconoscere con prudenza documenti amministrativi italiani, con OCR, catalogo tracciabile e funzionamento offline.

    Python 1

  4. building-agentic-automl building-agentic-automl Public

    Building an Agentic AutoML system from scratch, step by step, from a simple baseline to an experiment-driven senior ML agent.

    Jupyter Notebook 3

  5. building-llm building-llm Public

    A step-by-step educational journey from a character-level statistical language model to a small decoder-only Transformer.

    Jupyter Notebook 7 1

  6. home-credit-default-risk home-credit-default-risk Public

    Machine learning project to predict credit default risk with feature engineering, XGBoost and SHAP interpretability.

    Jupyter Notebook 6