Xray AI Lab
Learn how to evaluate LLM responses with test sets, code-based graders, human review, and LLM-as-a-judge, plus a runnable Python harness and common pitfalls.
Learn how backpropagation works: the chain rule, a fully worked numeric example, a NumPy implementation with gradient checking, and common pitfalls.
Learn a repeatable pandas data cleaning workflow: audit, fix text and types, parse dates, handle invalid values, remove duplicates, flag outliers, and validate.
Learn to measure data quality across six dimensions, detect and fix bad data with pandas, enforce rules with pandera, and see how bad data hurts models.
Run a complete exploratory data analysis in Python with pandas and seaborn on the Palmer Penguins data, from data quality checks to a Simpson's paradox.
Learn the probability behind machine learning: Bayes' theorem, key distributions, MLE and MAP, with worked examples and Python code you can run and check.
What is computer vision? Learn the core tasks, the key models from CNNs to Vision Transformers, real applications, and runnable Python examples.
Learn how gradient descent minimizes a loss function. Derive the gradients, code it in NumPy, and test learning rates with a plotted experiment.
Learn linear algebra for machine learning: vectors, dot products, matrix multiplication, and eigenvalues, with runnable NumPy examples and PCA.
Learn real-time object detection with YOLO in Python: bounding boxes, confidence scores, video inference, and tracking objects across frames.
Learn image segmentation with Segment Anything (SAM) in Python: point and box prompts, masks, bounding boxes, automatic mode, and honest limitations.
Vector RAG vs Graph RAG compared: what the papers found, the trade-offs, and a small reproducible Python experiment on multi-hop retrieval.
Learn how vision-language models work: image encoders, CLIP embeddings, captioning, and visual question answering, with a runnable Python demo.
Learn how machine learning detects cyber threats, from supervised models to anomaly detection, with a runnable Python example and honest limits.
Learn how Retrieval-Augmented Generation (RAG) works, why it outperforms basic LLMs for custom data, and how to build a working RAG app in Python.
Learn how AI agents work, explore real-world applications, compare top frameworks, and build a simple AI agent with Python.
Learn to build an AI agent with Python step by step. Set up your environment, implement tool calling, run the project, and troubleshoot common errors.
Learn data cleaning with Python and pandas: fix text, dates, types, duplicates and missing values with a tested pipeline you can check against ground truth.
Learn Python for data science from scratch: core syntax, NumPy, Pandas, Matplotlib, Seaborn, scikit-learn, a student scores project, and a learning roadmap.
Learn Python with worked examples: functions, lambdas, recursion, classes, number problems, series, and patterns, plus 28 graded exercises with solutions.
Understand convolution and pooling with runnable NumPy examples, then build, train, and evaluate a CNN image classifier in PyTorch (with a Keras version).
