Courses from 1000+ universities
AI got cheap enough that Duolingo’s most expensive plan may not survive it. I read the earnings call transcript and opened the app to see what is actually changing for learners.
600 Free Google Certifications
Computer Science
Artificial Intelligence
Data Science
Intelligenza Artificiale
The Bible's Prehistory, Purpose, and Political Future
Education for All: Disability, Diversity and Inclusion
Organize and share your learning with Class Central Lists.
View our Lists Showcase
Learn AI Models, earn certificates with free online courses from Stanford, University of Michigan, University of Virginia, CU Boulder and other top universities around the world. Read reviews to decide if a class is right for you.
Program Codex with the OpenAI Python SDK to build deterministic engineering workflows that safely plan, apply, and verify repository patches.
Evaluate multimodal AI systems with cross-modal metrics like FID, CLIP scores, and recall@k, then detect bias and apply interpretability and governance frameworks using LIME and SHAP.
Move AI projects from black box to business-ready: apply feature engineering, hyperparameter tuning, SHAP and LIME explanations, and structured experiments to fraud and credit models.
Fine-tune ViT-B/16 with transfer learning, calibrate predictions with temperature scaling, cluster anchor boxes for detection, and diagnose segmentation errors with IoU, Dice, and post-processing.
Build custom PyTorch layers, profile GPU bottlenecks, standardize ML workflows, and quantize models with TensorFlow Lite for deployment on Jetson Nano edge devices.
Evaluate model behavior across data slices and apply TensorFlow Lite quantization to weigh size, inference speed, and accuracy trade-offs before deploying to edge devices.
Customize pretrained models with transfer learning: apply supervised and instruction fine-tuning with LoRA in PyTorch and Hugging Face, audit fairness, and deploy monitored inference via FastAPI and Docker.
Build ensemble models with bagging, boosting, and stacking, weigh model complexity against interpretability, and validate improvements with paired t-tests, bootstrap resampling, and A/B experiments.
Adapt pretrained models to specific tasks: clean and tokenize datasets, run supervised and instruction fine-tuning, evaluate with F1, BLEU, ROUGE and perplexity, and apply LoRA and PEFT.
Prepare data with NumPy, Pandas, Matplotlib and Seaborn, build machine learning models, then adapt pretrained CNNs and explore transformer self-attention and encoder-decoder architectures.
Preprocess image data with pixel normalization and RGB, grayscale, HSV, BGR color-space conversions, then extract motion features from video using optical flow and frame differencing.
Design an AI operating model by balancing centralized, federated, and hub-and-spoke structures with explicit roles, decision rights, governance, and accountability.
Deploy fine-tuned AI models to production with FastAPI, Docker, and MLflow, applying quantization and distillation, monitoring drift and versions, and evaluating fairness with model cards.
Build a personal AI tutor in C# with DeepSeek by configuring model responses, conversation history, system prompts, and multiple chat sessions.
Build a personal AI tutor in PHP with DeepSeek by configuring API parameters, managing conversational history and sessions, and designing custom system-prompt personas.
Get personalized course recommendations, track subjects and courses with reminders, and more.