Tuan Linh Dao

Project 3) T-SQL Agent (in testing phase) A specialized SQL conversational agent that allows querying the SQL Server database by taking a natural language question as input and generating a corresponding T-SQL query. Tools used: SQL Server, Python (Numpy, FastAPI, Langchain-FAISS, SQLAlchemy), LLM (Mistral:7B pre-trained locally), and web development languages ​(HTML, JavaScript, CSS). Project 2) Predictive Modeling Project at Air Corsica Development of a customer purchase probability model using an exponential distribution adjusted by recency, seasonality, and recent return. The model is validated on rolling data (24 months of history) with an accuracy rate greater than or equal to 85% over 15 consecutive months. Tools used: SQL Server, Python (Numpy, SQLAlchemy, Pandas, Matplotlib). Project 1) Age and gender prediction from a person's first name Predicting gender and age from a person's first name allows for personalized offers and messages, adapting marketing campaigns to demographic profiles, even in the absence of explicit customer information. Models developed: Neural network Regression with PyTorch for age prediction, validated by an MSE of 0.38 Gender classification with TensorFlow, validated with an accuracy of 98% Tools used: SQL Server, Python (Numpy, Scikit-learn, Pandas, Matplotlib, TensorFlow, PyTorch)

This is my first internship in France. It is carried out from April 11 to June 19, 2022 in the LARHRA laboratory in Saint Martin d'Hères, Grenoble to validate my DUT diploma in Statistics and Business Intelligence. I joined the DYNAMO project on the analysis of the capacity of mountain sports professions in which I do the descriptive statistical analysis of the responses to the survey by questionnaire and segmentation of the responses of ski instructors.