Freelancer profile translated to English.

Description

Data Scientist and AI Engineer graduated with a double Master's degree in Data Science and Artificial Intelligence (Institut Polytechnique de Paris & AIT Bangkok).

I assist startups and companies in setting up cloud data architectures (Azure), developing machine learning/NLP models, and analyzing complex data to transform information into strategic decisions.

Experience:

• Design and deployment of a data processing architecture on Azure for a smart-city startup in Paris.

• Development of a fine-tuned translation model for a low-resource language during a research internship.

🔹 My strengths: autonomy, scientific rigor, creativity, and a sense of real impact.

Available for freelance missions in Data Science, AI, NLP, and Cloud Engineering.

Languages

English
Native or bilingual
French
Conversational
Burmese
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

Floware (École Polytechnique X Innovation Center Incubator)
Data Science/Cloud Data Engineer
July 2024 - June 2025 (11 months)
91300 Massy, France
• Engineered and deployed a scalable Azure-based data architecture integrating multi-source sensor, mobility, and geospatial datasets.
• Automated ETL workflows with PySpark, improving efficiency by 35%.
• Developed forecasting and anomaly detection models for urban mobility.
• Optimized cloud resource allocation, achieving 20% cost savings.
Python Data science Microsoft Azure Data analysis Machine learning
DiCE Lab
Assistant Researcher ( Natural Language Processing
August 2023 - July 2024 (11 months)
Paris, France
• Initiated and led a research project on low-resource language translation, using Python and Hugging Face Transformers.
• Collected, cleaned, and augmented datasets, boosting BLEU score by 18%.
• Produced reproducible research pipelines and internal documentation.
Python Microsoft Power BI
AIT BrainLab
Research Assistant ( Natural Language Processing)
January 2023 - August 2023 (7 months)
Bangkok, Thailand
• Designed a paraphrasing tool leveraging sequence-to-sequence models, improving output fluency.
• Streamlined preprocessing pipelines, raising accuracy by 15% and reducing prep time by 25%.