AI classes built by AI. The good kind.
Prompt engineering, agents, LLM internals, AI products, pick a topic and Octo writes a class shaped around what you do for a living.
Building AI Products with LLMs
Your demo works on three prompts and falls apart on the fourth. Learn the patterns that make LLM features actually hold up once real users start typing into them.
RAG From Scratch
Stop copy-pasting RAG tutorials that hallucinate citations. Build retrieval, chunking, hybrid search, and evals end to end, so you understand why each piece is there.
LLM Evals & Testing
You shipped an AI feature and have no idea if it is getting better or worse. Learn to build golden sets, offline evals, and regression suites so you can ship with confidence.
Production AI Systems
Your prototype costs ten cents per call and takes eight seconds. Learn the caching, fallback, and observability patterns that turn AI demos into systems that scale.
Fine-Tuning & Distillation
Stop guessing whether to fine-tune or just write a better prompt. Learn when each approach actually wins, with hands-on LoRA, SFT, and distillation you can apply this week.
Multimodal AI Engineering
Your users send screenshots, voice notes, and PDFs, not clean text. Learn to build pipelines that handle vision, audio, and documents without falling apart at the seams.
Your First AI Agent
You have read about agents and watched the demos, now build one yourself. Go from a single tool call to a multi-step working agent you actually understand.
Agent Architectures in Production
Most agent demos die the moment a real user uses them. Learn the architectures that hold up, single-agent, multi-agent, supervisor-worker, and how to pick the right shape.
Tool Use & Function Calling
Your agent picks the wrong tool, hallucinates arguments, and silently fails. Learn to design schemas, handle errors, and parallelize calls so your tools actually work.
MCP & Agent Protocols
MCP is suddenly everywhere and the docs assume you already get it. Learn the Model Context Protocol and the emerging agent stack from the ground up.
Agent Evals & Safety
You cannot put an agent in production if you cannot trust what it does. Learn sandboxing, permissions, jailbreak resistance, and the eval methods that catch real failures.
Prompt Engineering for Real Work
Stop writing prompts that kind of work and hoping for the best. Learn the structure, patterns, and iteration loop that turn AI from a toy into a tool you trust.
Advanced Prompting Techniques
You have outgrown basic prompting and your output still feels random. Learn chain-of-thought, ReAct, self-consistency, and routing, so you finally get reliable behavior.
Prompt Engineering for Coding
Cursor and Claude Code keep generating code you have to throw away. Learn the prompting patterns that get useful diffs the first time, not the fifth.
Prompting for PMs & Marketers
Your AI drafts read like a robot wrote them, because one did. Learn the prompting moves that produce specs, copy, and research synthesis you can actually ship.
Structured Output & Function Calling
You need clean JSON and the model keeps giving you prose. Master schemas, function calling, and structured output so your AI feature plays nicely with the rest of your code.
Generative AI for Creators
There are six new generative tools every month and most of them are toys. Learn what Midjourney, Sora, ElevenLabs, and Suno are actually good at, and how to direct them.
Image Generation Deep Dive
Move from lucky prompts to consistent, on-brief images. Learn how diffusion models work, the control techniques that actually steer them, and the rules around commercial use.
Video Generation & Editing
AI video is finally usable for client work, if you know what you are doing. Master Sora, Veo, and Runway with workflows that survive a real production timeline.
Voice & Audio AI
Voice and audio AI moved from novelty to product overnight. Learn voice cloning, music generation, and conversational audio, plus the ethics and consent rules nobody talks about.
GenAI for Marketing
Your team is generating content that all sounds the same and looks like a stock photo. Learn the briefs, ad creative, and lifecycle copy workflows that hold brand quality.
Machine Learning Fundamentals
Most ML courses bury you in math before you touch a model. Learn linear models, trees, ensembles, and evaluation in a way that actually sticks, with the math kept in service of the ideas.
Deep Learning From Scratch
You can use a PyTorch model and you have no idea what it is doing. Build backprop, CNNs, RNNs, and transformers up from scratch, so the architecture stops being a black box.
MLOps & Model Deployment
You trained a model and now nobody knows what to do with it. Learn the serving, monitoring, and retraining work that takes ML from notebook to production.
Feature Engineering & Data Prep
Models get the credit but features do the work. Learn the data prep, leakage checks, and reproducibility patterns that decide whether your model is any good.
Reinforcement Learning Applied
RL went from academic to mandatory the day RLHF made GPT useful. Learn the fundamentals plus RLHF and RLAIF, so you understand the alignment work everyone is talking about.
How LLMs Actually Work
You use LLMs every day and have only a hand-wavy sense of what they are. Learn tokens, attention, training, and alignment clearly, with the math kept to what you actually need.
Transformers From Scratch
You can read a transformer paper and you cannot quite implement one. Build a small transformer end to end, so every line stops being a mystery and starts being a choice.
RLHF & Model Alignment
How does a next-token predictor become a useful assistant. Learn RLHF, DPO, and the alignment pipeline, plus the failure modes that explain weird model behavior in the wild.
Open-Source LLM Stack
You want to stop paying per token and own your stack. Learn to choose, deploy, and fine-tune Llama, Mistral, and Qwen, so open-source LLMs become a real option for your team.
Long Context & State
A million-token window sounds magic until you put real data in it. Learn the context engineering and memory patterns that make long context work, plus the trade-offs nobody tells you about.
AI Product Management Foundations
You are PM-ing your first AI feature and your old PRD template is useless. Learn how AI products differ, and how to plan, scope, and ship them without flying blind.
AI Feature Discovery & Scoping
Every team is being told to add AI and most of those features will flop. Learn where AI genuinely adds value, where it adds risk, and how to tell which is which before you build.
Evals for PMs
You do not need to write the model code, but you do need to know if it is working. Learn how to design and read evals as a PM, so you can lead the AI feature instead of trusting it.
AI PM Interview Prep
AI PM interviews ask things your old loop never did. Learn to design an AI feature, define an eval, plan a rollout, and manage risk on the spot, with rubrics that match real loops.
Responsible AI for Product Teams
Responsible AI usually means a slide deck that helps no one. Learn the bias, transparency, and consent work that actually changes what you ship, plus the EU AI Act in plain English.
Computer Vision Fundamentals
You want to ship vision features and the field changes every quarter. Learn CNNs, detection, and segmentation cleanly, so the modern CV toolkit stops feeling like an alphabet soup.
Vision-Language Models
CLIP, BLIP, and GPT-4V suddenly do things classic CV never could. Learn how vision-language models work and exactly what they unlock for your product.
Object Detection in Practice
You can train YOLO in Colab and you cannot ship it on a real device. Learn the deployment trade-offs that decide whether your model survives once it leaves the GPU.
Document Understanding
Your business runs on PDFs and your code cannot read them. Master OCR, layout parsing, and table extraction, so messy documents finally become structured data.
Modern NLP Foundations
Modern NLP is not the NLP your textbook describes. Learn tokenization, embeddings, and transformers as they actually work today, so you can build past chatbots.
Semantic Search with Embeddings
Keyword search returns garbage and your users notice. Learn vector search, hybrid retrieval, and reranking, the same stack that powers any RAG system worth using.
Text Classification at Scale
You are paying an LLM to do work a small model could do for a fraction of the cost. Learn when a tiny classifier beats an LLM, and how to scale either one when traffic grows.
Information Extraction
Your team spends days copy-pasting fields out of emails and PDFs. Learn to extract entities, relations, and events from messy text, so the work stops being manual.