TXT Group is an international, IT end-to-end provider of consultancy, software solutions and services, supporting the digital transformation of customers’ products and core processes. With a proprietary software portfolio and deep expertise in vertical domains, TXT Group operates across different markets, with a growing footprint in Aerospace, Aviation, Defense, Industrial, Government and Fintech. The holding company, TXT e-Solutions, has been listed on the Italian Stock Exchange - STAR segment (TXT.MI) - since July 2000. TXT Group is headquartered in Milan and has subsidiaries in Italy, Germany, the United Kingdom, France, Switzerland and the United States of America.
**Thesis Project - Audio Signal Processing for Aviation**:
**Project Description**:
The goal of the project is to design and develop a system for the acquisition, processing, and recognition of audio signals generated within the cockpits of airplanes and helicopters. The audio signals of interest include both voice commands (even complex phrases) and simple acoustic signals, such as beeps.
Audio can be acquired either through direct digital connections (e.g., cables) or via environmental microphones, which entails significant background noise exposure, such as machinery sounds, electrical disturbances, conversations, and human movements.
The system will be designed to address the following challenges:
**Audio preprocessing**: denoising, volume normalization, equalization, and other transformations to improve signal quality;
**Variability management**: dynamic handling of playback speed, voice timbre,



and pitch based on configurable parameters;
**Classification vs. matching**: ability to operate on a finite dataset of known sounds or compare acquired signals with generic templates in an open-set approach;
**Post-processing**: refinement of the output to adapt it to other systems or to enhance readability.
The dataset will be built specifically for the project, combining synthetic and real sounds with controlled variants. Technologies and libraries such as Python, librosa, PyTorch, and TensorFlow will be used, along with classical signal processing techniques (e.g., Fourier Transform, MFCC) and neural approaches (CNN, RNN, transformers).
**Recommended Skills**:
Basic knowledge of audio signals (amplitude, frequency, spectrum);
Familiarity with transforms (Fourier) and signal processing techniques;
Fundamental understanding of machine learning;
Proficiency in Python and major data science libraries.
**We Offer**:
Internship opportunity with the possibility of future integration into the company and career growth in an international environment;
Flexible work environment, including flexible working hours and lavoro da remoto policies;
**International exposure**: global presence with offices in Milan, Berlin, Seattle, Dallas, Paris, and daily collaboration among multicultural teams;
Strong team-oriented culture, with team-building activities and corporate events;
**Internship location**: possibility to work in a hybrid mode from our Cologno Monzese office.
Li-Hybrid

📌 Thesis Project - Audio Signal Processing for Aviation (Città Metropolitana di Milano)
🏢 TXT e-solutions
📍 Città Metropolitana di Milano