Freelancer profile translated to English.
Back to original languageAbout Duc-Trong
Do you need to leverage your data, automate Python processes, or build a reliable predictive model?
I can assist you with data science, machine learning, and data analysis missions, from exploration to the production of clear and actionable deliverables. I handle issues related to predictive modeling, classification, regression, automation, model optimization, and data pipeline structuring.
My approach is analytical, rigorous, and results-oriented. I don't just train a model: I analyze data, test various approaches, compare performance, improve variables, and seek the most relevant solution for your business needs.
I can help you with:
data analysis and cleaning, Python scripts, feature engineering, model benchmarking, visualization, analysis reporting, and developing robust and reusable pipelines.
Possible deliverables: Python scripts, analysis notebooks, predictive models, processing pipelines, visualizations, summary reports, and documentation.
French
Native or bilingual
English
Native or bilingual
Vietnamese
Native or bilingual
Can work on-site
Toulouse (up to 50km)
Experience
- Saigon ImgFreelance – Geospatial Image Classification / Computer VisionAVIATION AND AEROSPACEOctober 2025 - January 2026 (3 months)Toulouse, FranceCarried out a geospatial image classification mission with the development of a complete computer vision pipeline, from data preparation to model evaluation.- Collection, sorting, and organization of satellite/aerial images into several landscape classes- Dataset cleaning, image quality control, and class rebalancing to improve training stability- Image preprocessing: resizing, normalization, label encoding, and implementation of data augmentation techniques- Development and training of deep learning CNN models for automatic image classification- Monitoring of performance on training and validation sets, comparison of configurations, and hyperparameter tuning- Detailed analysis of prediction errors to identify confusions between visually similar classes and improve model robustness- Structuring a reusable workflow to facilitate retraining, testing new architectures, and comparative evaluation of resultsTechnologies: Python, PyTorch, Computer Vision, CNN, Data Augmentation, NumPy, Pandas, TensorBoard
- TDF (Télédiffusion de France)Data ScientistTELECOMMUNICATIONSMarch 2025 - September 2025 (6 months)Metz, FranceAs part of a predictive modeling project applied to radio wave propagation, I worked on a dataset of 360,000 rows and 70 variables to evaluate, compare, and optimize several machine learning approaches.I calibrated and analyzed 8 models using statistical indicators, metrics like RMSE and MAE, and multi-criteria validations. I also developed and optimized several machine learning models (linear regression, Random Forest, XGBoost, MLP), achieving accuracy improvements of up to 40%.I designed an end-to-end automated pipeline covering data cleaning, feature engineering, variable selection, training, validation, and reporting, which reduced training time by 30% and ensured experiment reproducibility.Finally, I participated in the industrialization of the developed solutions by exporting Python models to C++ for integration into Atoll.Technologies: Python, Scikit-learn, TensorFlow, XGBoost, Pandas, NumPy, C++, Git, Bash, Xarray, Flask
- Vietnam Academy of Science and TechnologyData AnalystApril 2024 - August 2024 (4 months)I worked on a climate data analysis project focused on precipitation in Southeast Asia, using time series data from CNRS climate models and observational data.I performed data processing, statistical analyses, and comparisons between models and observations to evaluate discrepancies, uncertainties, and the reliability of projections. I also applied several bias correction methods (EQM, SCL, LOCI) to improve the accuracy of climate projections.This mission allowed me to strengthen my skills in data analysis, statistics, time series processing, and complex data validation.Technologies: Python, Pandas, NumPy, Xarray, NetCDF, CDO, Ferret
Recommendations
Be the first to recommend Duc-Trong
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master Applied Mathematics for EngineeringUniversité Paul Sabatier2025Master Mathématiques appliquées pour l'ingénierie