AIMachine Learning Intern

Thales is a global technology leader trusted by governments institutions and enterprises to tackle their most demanding challenges. From quantum applications and artificial intelligence to cybersecurity and 6G innovation our solutions empower critical decisions rooted in human intelligence. Operating at the forefront of aerospace and space cybersecurity and digital identity were driven by a mission to build a future we can all trust.

TOPIC : Speech Denoising

Description

This internship offers the opportunity to contribute to R&D on advanced AI Speech Denoising initiative aimed at improving speech clarity in real-world noisy environments. The intern would combine deep learning digital signal processing and data-driven optimization to build robust models capable of separating human speech from complex background noise. The initial focus would be on specializing and fine-tuning state-of-the-art noise suppression AI models to achieve optimal denoising performance in selected acoustic environments.

Responsibilities:

• Data Collection & Preparation
• a) Collect annotate and preprocess speech and noise datasets from various environments.
• b) Implement scripts for mixing speech and environmental noise at different signal-to-noise ratios (SNRs) and perform quality control on generated data.
• Machine Learning Development
• a) Assist in training and fine-tuning deep learning models for speech denoising.
• b) Evaluate models using objective metrics (e.g. PESQ STOI) and subjective listening tests.
• c) Experiment with architecture variations training configurations and hyperparameter tuning.
• System Integration & Evaluation
• a) Conduct performance benchmarking and error analysis.
• b) Document results visualize findings and present outcomes to the research team.

Requirements:

• Education: Currently pursuing or recently completed a Bachelors or Masters degree in Computer Science Electrical Engineering or a related technical discipline.
• Programming: Proficient in Python with experience in NumPy pandas and preferably PyTorch.
• Machine Learning Knowledge: Familiar with deep learning fundamentals and general model training workflows.

Expected Outcomes

• Hands-on experience building and evaluating state-of-the-art speech denoising systems.
• Exposure to end-to-end AI development from data acquisition to deployment.
• Mentorship from researchers and engineers experienced in audio AI deep learning and signal processing.
• A practical understanding of how AI models integrate into real-world audio pipelines.

Commitment:

• Able to commit on a full-time basis for ideally 6 months (or a minimum of 5 months) from Jan 2026.

At Thales were committed to fostering a workplace where respect trust collaboration and passion drive everything we do. Here youll feel empowered to bring your best self thrive in a supportive culture and love the work you do. Join us and be part of a team reimagining technology to create solutions that truly make a difference for a safer greener and more inclusive world.