Machine Learning Reply Italy

Cristian Cosci

Senior Machine Learning Engineer

  • Generative AI
  • LLMs
  • Computer Vision
  • Optimization

About

I'm a Senior Machine Learning Engineer at Machine Learning Reply in Turin, where I design and build AI solutions in Generative AI and LLMs, Computer Vision, and Machine Learning and optimization for logistics.

Before joining Reply I was a Researcher in Artificial Intelligence at the University of Bologna and University of Pisa, working on Generative AI and Trustworthiness. I've given several talks on AI and AI applied to logistics.

Experience

  1. Machine Learning Reply

    Turin, Italy

    1. Senior Machine Learning Engineer Oct 2026 – Present
    2. Machine Learning Engineer Dec 2024 – Oct 2026
    • Generative AI & LLMs: adapting open-source LLMs to enterprise domains through Continual Pre-Training and Fine-Tuning, from data preprocessing and synthetic data generation to evaluation
    • Computer Vision: design and development of Computer Vision solutions for enterprise use cases
    • Logistics & Optimization: Machine Learning and optimization algorithms for intermodal rail and road transport planning and for the scheduling of container terminal operations

  2. University of Bologna and University of Pisa

    Bologna and Pisa, Italy

    1. Researcher in Artificial Intelligence Nov 2023 – Nov 2024
    • Research on Generative AI and Trustworthiness

  3. University of Perugia

    Perugia, Italy

    1. Research Intern Sep 2022 – Jan 2023
    • Adversarial attacks on image classification models that exploit explainability methods based on activation maps

Publications

Google Scholar

  1. 2025

    • ECAI 2025
    • CORE A

    Pirates of the RAG: Adaptively Attacking LLMs to Leak Knowledge Bases

    C. Di Maio, C. Cosci, M. Maggini, V. Poggioni, S. Melacci

    Proceedings of the 28th European Conference on Artificial Intelligence (ECAI 2025), Frontiers in Artificial Intelligence and Applications, vol. 413, pp. 4041–4048

Featured projects

All repositories

  • complex-network-link-prediction

    Python library, published on PyPI, implementing the main link-prediction techniques for complex networks

    Complex networks Python

  • Ant Colony Optimization for OSSP

    Open Shop Scheduling Problem solved with Ant Colony Optimization

    Optimization Python

  • LLMs Robustness Evaluation

    Robustness evaluation of LLMs on sentiment analysis

    LLMs Jupyter Notebook

  • GAN Overview

    Overview of GANs and their main variants, with implementations

    Generative AI Python

  • AOPC MoRF

    Area Over Perturbation Curve (Most Relevant First) to evaluate XAI methods

    Explainable AI Python

  • GMRES

    GMRES for the numerical solution of indefinite non-symmetric linear systems

    Numerical analysis MATLAB

Talks, Teaching & Service

  • Speaker at several events on AI and AI applied to logistics

  • Program Committee Member

    2nd AEQUITAS Workshop on Fairness and Bias in AI, co-located with ECAI 2024

    2024

  • Teaching Assistant

    Machine Learning with Laboratory (M.Sc.), University of Perugia

    Feb 2024 – Jun 2024

  • Teaching Assistant

    Procedural Programming with Laboratory (B.Sc.), University of Perugia

    Nov 2022 – Feb 2023

Skills

  • Generative AI

    • Continual Pre-Training
    • Fine-Tuning (LoRA)
    • RAG
    • synthetic data generation

  • Optimization

    • Operations Research
    • Constraint Programming with Google OR-Tools (CP-SAT)

  • Machine Learning

    • PyTorch
    • TensorFlow
    • Keras
    • scikit-learn
    • OpenCV
    • NumPy
    • SciPy
    • Pandas
    • NetworkX

  • Cloud & tools

    • AWS
    • GCP
    • Azure
    • Docker
    • Linux
    • Git
    • LaTeX

  • Programming

    • Python
    • C
    • Java
    • MATLAB

  • Languages

    • Italian (native)
    • English (C1)

Education

  1. 2021 – 2023 110/110 cum laude

    M.Sc. in Computer Science, Artificial Intelligence curriculum

    University of Perugia

    ThesisAdversarial Attacks on Explainability Methods for Image Classification Models

  2. 2018 – 2021 110/110 cum laude

    B.Sc. in Computer Science

    University of Perugia

    ThesisBitcoin price-trend prediction using Machine Learning

Schools

  • ESSAI & ACAI 2024, Athens
  • AI & Society Summer School 2024, Tropea (Best Project Award)
  • School in AI 2022, UniMoRe
  • CyberChallenge 2020