Flashcards.
Active recall on the ideas that matter — scheduled by an SM-2 algorithm so each card comes back right before you'd forget it. Grade yourself; the harder ones return sooner. Progress saves in your browser.
The Agent Loop & Harness
Plan-act-observe, tool calling, context management — the runtime behind every AI agent.
Loop Engineering
Outer loops, stopping conditions, verification splits — running agents without being the keyboard.
Transformers & Attention
Tokens, attention, and the architecture under every modern LLM.
System Design Foundations
Caching, scaling, consistency — the load-bearing ideas of every large system.
LLM Engineering
Decoding, RAG, evals, and cost — building reliable products on top of models.
Context Engineering
The attention budget, the four operations, compaction, caching — what the model sees and why.
Inference & Serving
KV cache, batching, quantization, TTFT — how tokens actually get made, fast and cheap.
RAG & Retrieval
Chunking, hybrid search, reranking, grounding — the retrieval half of every AI product.
Landmark AI Papers
One-line recall for the papers everyone cites — from Attention to R1.
Algorithms & Data Structures
The core interview toolkit — complexity, the right structure for the job, and the patterns behind most problems.
AI Safety & Security
The new attack surface and the defenses — prompt injection, jailbreaks, guardrails, and evaluating for harm.
The FDE Interview
The nine round archetypes and what each one actually scores — decomposition, learning, incremental coding, the defended take-home.
FDE Practice: Deployment to Renewal
Scoping, evals, identity, on-call and the commercial layer — the facts that decide whether a deployment survives.
Glossary: AI & LLMs
Every AI and LLM term in the glossary, as recall pairs — models, training, agents, evaluation.
Glossary: Systems & Backend
The systems vocabulary — consistency, queues, caching, failure modes — straight from the glossary.
Glossary: Data & Retrieval
Embeddings, indexes, chunking, ranking — the retrieval half of the glossary as flashcards.
How it works: each card gets an ease score. Rate a card Again / Hard / Good / Easy and the scheduler picks when you'll see it next — failed cards return tomorrow, easy ones drift weeks out. Come back daily and the deck teaches itself. Pairs with your daily streak.