Prompt-Based AutoML Assistant for Non-Technical Users

Many professionals — especially in healthcare, education, and agriculture — are surrounded by valuable data but lack the technical know-how to apply machine learning. Traditional AutoML tools either require technical expertise or lack real-world contextual understanding.

This talk introduces Prompt-Based AutoML Assistant — a free and open-source project that enables users to build entire ML workflows simply by typing plain English prompts like:

“Predict diabetes risk using age, BMI, glucose, and blood pressure.”

This assistant automatically:

• Understands the domain from the prompt

• Accepts datasets via upload/API

• Cleans and analyzes the data

• Trains optimal models

• Applies SHAP for explainability

• Outputs an interactive dashboard or API

Built with Python, GPT APIs, Scikit-learn, XGBoost, LangChain, and Streamlit, this modular tool is designed to democratize AI, especially for non-technical users. It’s containerized with Docker, plug-and-play, and perfect for rapid deployment in real-world domains.

• Understand the architecture behind prompt-to-ML pipeline systems

• Learn how to combine LLMs with AutoML for real-world applications

• Explore how open-source tools like Streamlit, Scikit-learn, LangChain, and Docker can be used together effectively

• Discover how to build ML tools for non-technical users

• See a live demo of a real-world healthcare use case