Meyssa User

I’m an Applied Machine Learning Engineer with a Master’s in Artificial Intelligence and 4+ years building enterprise data and software systems. I focus on classification and evaluation rigor—especially feature engineering, imbalanced data, threshold tuning, and hyperparameter tuning—so models translate into reliable decisions.

In my current role, I build applied ML workflows for faculty-level course enrolment prediction using Azure AI Foundry, Microsoft Fabric, and Azure AutoML. I engineer ML-ready datasets from Oracle sources with cleaning, validation, feature engineering, stratified sampling, and historical pipelines, then address imbalanced classification with weighted loss functions, threshold tuning, and model benchmarking.

Earlier, I developed and maintained PeopleSoft applications and integrations, including REST and SOAP connectivity and SQL optimization to improve performance and reporting accuracy. I also prototype RAG-assisted text-to-SQL (SchemaAgent), hybrid retrieval for question answering (Quora-RAG), and interpretable LightGBM models using Optuna and SHAP.