Need a machine learning model that actually works, not just fits your training data? I build regression and classification models using Python, scikit-learn, and LightGBM, with a focus on results you can explain to stakeholders, not black boxes.
Recent work: an end-to-end risk classification system with explainable AI (SHAP-style interpretability), and a regression model predicting building energy loads from structural data. I also work with image data for computer vision tasks using PyTorch, and I'm actively expanding into LLM-based systems (RAG, AI agents with LangChain).
What you get: a working model, a clear explanation of the trade-offs I made (accuracy vs. interpretability, precision vs. recall), and visualized results (Matplotlib/Seaborn) so you understand why it works, not just that it does.
Message me with your dataset and goal. I'll tell you honestly whether it's a quick job or a bigger build before we start.