I built and benchmarked predictive models for real estate prices using the Divar Real Estate Ads Dataset. I compared Neural Network, Random Forest, and XGBoost models with cross-validation, and used K-Means to segment the market.
For the same project, I created a Persian ad classification pipeline using Hazm, CountVectorizer, and Word2Vec. I evaluated the models with precision, recall, and F1-score, then translated model output into price-driver insights for non-technical stakeholders.
My statistical modeling work includes regression and causal inference analyses in R on social and economic datasets, as well as climate data analysis in Python and Xarray. I also built a RAG-based question-answering pipeline over PDF documents using embeddings, ChromaDB, and semantic retrieval.

