Freelancer profile translated to English.

Description

AI Specialist, I help companies leverage machine learning and LLMs to automate, predict, and optimize. I transform your ideas into concrete, scalable, and tailor-made solutions. Experienced in managing projects from A to Z, from proof of concept to production deployment.

Languages

French
Native or bilingual
English
Fluent

Workplace preferences

Remote only

Primarily works remotely

Sphere
Data Scientist
CONSULTING AND AUDITS
March 2026 - Today (3 months)
Paris, France
B2B application development (Backend + AI Features):
- Generation/Deployment of AI content
Python API Artificial intelligence SQL Prompt Engineering
AUBAY
AI Research Engineer Intern
DIGITAL AND IT
February 2025 - Today (1 year and 4 months)
Paris, France
This internship is part of applied research aimed at using language models (LLMs) and signal processing algorithms to automatically detect and correct anomalies in audio files.

1. State of the Art

- Study of audio anomaly detection methods via Voice Activity Detection (VAD) and other signal processing techniques.

- Research and comparison of LLM model architectures suitable for audio analysis (Hugging Face, Whisper, Wav2Vec2, etc.).

- Experimentation with various audio segmentation and feature extraction methods (MFCCs, spectrograms, zero-crossing rate, etc.).

2. Proof of Concept (PoC) Development – Algorithms & LLMs

- Design and implementation of a Python-based detection pipeline: Using PyAudio, Librosa, torchaudio for audio signal preprocessing.

- Development of a speed variation tracking algorithm (WPM - Words Per Minute) to identify accelerations/decelerations.

- Application of Deep Learning & NLP models to classify detected anomalies.

- Model optimization via fine-tuning on specific datasets.

- Comparative analysis of model performance (precision, recall, F1-score) across different learning configurations.

3. Deployment & Validation – WebApp and real-world tests

- Development of a user interface in Angular, containerized with Docker.
Implementation of two usage modes:

- Live mode: real-time error detection with anomaly type notification.

- Batch mode: upload audio files, classification, and automatic correction of detected anomalies.

- Integration of an Audio-to-Text module to validate the impact of corrections on transcription quality.

4. Project Management & Collaboration

- Use of GitLab/GitHub (CI/CD, merge requests).

- Task tracking and organization in Agile sprints via Jira.
Artificial intelligence Python Deep Learning Natural Language Processing (NLP)
Mairie de Malatya
AI-Cyber Engineer Intern
DEFENSE AND MILITARY
June 2024 - August 2024 (3 months)
Malatya, Türkiye
Mandatory 3-month internship in the web and telecommunications departments of the Battalgazi municipality.

Missions:

- AI project for anti-cyber attack alarm: Machine learning development of an AI to detect virtual attacks (DDOS, phishing, etc.) on the server and notify the telecommunications service.

- Management of .csv files and creation of dashboards for the anti-cyber attack alarm AI project

- Development of a chatbot on the website to redirect users to the correct tabs.

- Learning C# and ASP.NET CORE used for the development of the municipality's website.

- Refresher on SQL and JavaScript languages

Skills used:

Python
Machine Learning
C#
ASP.NET CORE 7.0
SQL
JAVASCRIPT
Cybersecurity and Networks (different forms of attacks like DDOS and defense mechanisms including firewalls)
Python Artificial intelligence SQL Cybersecurity