← All drills

I want to build an app that answers from my own documents

Retrieval, prompting and checking that the answers are right.

Skills this builds

Retrieval, RAG and vector searchLarge language modelsPrompt engineeringHugging FaceModel evaluation

Start here · beginner and intermediate

Build a Text-to-SQL Agentintermediate
Write a test set whose answers are SQL, and grade an agent by comparing the rows its SQL returns
5 sections · Snowflake, Databricks, Google

Covers 4 of 5 skills: Large language models, Model evaluation, Prompt engineering, Retrieval, RAG and vector search

Build a Document Search Engineintermediate
Build an inverted index and answer keyword queries over thousands of documents
4 sections · Elastic, Algolia, Apache Lucene

Covers 2 of 5 skills: Model evaluation, Retrieval, RAG and vector search

Build a Next-Word Predictor from Scratchbeginner
Build a predictive-text engine from scratch — the next-word suggestions behind a phone keyboard
8 sections · Google, Apple, Microsoft

Covers 2 of 5 skills: Large language models, Model evaluation

Build a ReAct Agentintermediate
Build the ReAct loop from scratch in pure Python — reason, act, observe, repeat
9 sections · OpenAI, Anthropic, LangChain

Covers 2 of 5 skills: Large language models, Prompt engineering

Build an Agent Memory Engineintermediate
Build long-term memory for AI agents
10 sections · Maximem, Mem0, Zep

Covers 2 of 5 skills: Large language models, Retrieval, RAG and vector search

Build LLM Applications with LangChainintermediate
Build a chain of a prompt template, a chat model and an output parser, and measure how reliably it returns a clean answer
9 sections · Klarna, Rakuten, Replit

Covers 2 of 5 skills: Large language models, Prompt engineering

Build Your First Agent: Thought, Action, Observationintermediate
Build an LLM-powered agent
5 sections · Intercom, Zendesk, GitHub

Covers 2 of 5 skills: Hugging Face, Large language models

Fine-Tune GPT-2 to Tag Legal Contract Clausesintermediate
Fine-tune GPT-2 on real legal contract clauses so it tags a clause with its type
9 sections · Harvey, Casetext, LexisNexis

Covers 2 of 5 skills: Hugging Face, Large language models

Test Your Prompts Like Codeintermediate
Treat a prompt as testable data, and run it over a labelled set of cases with a harness you wrote
11 sections · OpenAI, Anthropic, Google

Covers 2 of 5 skills: Model evaluation, Prompt engineering

Go further · advanced