Gradient Descent and Optimizer Geometry: Momentum, Adam, and Loss Surfaces
Hand-calculate early steps on a 2D quadratic, compare gradient descent, momentum, and Adam, and reproduce the loss-contour visualization.
AI Learning Project
Convolution and Receptive Field Math: Padding, Stride, Channels, and im2col
A hand-calculated 5×5 by 3×3 convolution with output-size math, receptive fields, channel mixing, and im2col.
Transformer Attention Math: Q/K/V, Softmax Weights, Masks, and KV Cache
A three-token hand calculation of scaled dot-product attention, covering Q/K/V, softmax weights, masks, multi-head attention, and KV cache.
Matrix Calculus for Neural Networks: Deriving the MSE Gradient
A visual and reproducible derivation of dL/dW = (y_hat – y)x^T for y = Wx + b, with NumPy gradient checking.
Backpropagation as a Computation Graph: A Two-Layer MLP by Hand
Trace a two-layer MLP through local gradients, ReLU, softmax cross-entropy, and a reproducible NumPy backward pass.
LLM, RAG, and Agent Security: Prompt Injection, Tool Permissions, and Boundary-Aware Defense
A practical guide to LLM/RAG/Agent security covering prompt injection, external content isolation, tool allowlists, approval gates, and boundary-aware defenses.
Data Poisoning and Backdoor Defense: Poison Rates, Triggers, and Training Pipeline Isolation
A defensive guide to data poisoning and backdoors, with a toy digits experiment, poison rate, attack success rate, and training pipeline controls.
Model Privacy and Extraction Defense: Membership Inference, Surrogates, and Prediction API Controls
A practical guide to model privacy risks: membership inference, surrogate extraction, confidence leakage, output minimization, and query governance.
AI Security Threat Modeling: Build a Defense Map with NIST, MITRE ATLAS, and OWASP
A practical AI security threat modeling guide using NIST adversarial ML, MITRE ATLAS, and OWASP LLM Top 10 to map assets, risks, evidence, and controls.
Adversarial Examples and Robust Evaluation: From FGSM to a scikit-learn Digits Experiment
A defensive robustness evaluation walkthrough using the FGSM equation and a safe scikit-learn digits toy experiment.