Built for engineers, ML & data teams

Stop bookmarking.
Start scouting.

A Scout Loop is a sharp 40-minute first pass on any tool or idea: real context, one practical action, and a decision at the end. No more open-ended tab graveyards.

Request access →See a Scout Loop

Invite-only — early access open

See one before you sign in

Sample Scout Loops. Open to everyone.

Real, source-grounded examples — not a course catalog. Each is a 40-minute scouting sprint you could run today.

Scout Loop · learn it in a 40-minute sprint

Agent Observability / Tracing Sprint

For technical operators

Agent observability is the habit of treating an agent run as a workflow you can inspect, not just a blob of prompt text and final output. In this Scout Loop, you create one tiny Python agent with a tool, run it inside a named trace, and inspect whether the trace shows the workflow shape you expected.

bash

python3 --version
mkdir -p agent-trace-scout
cd agent-trace-scout
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip openai-agents

Scout Loop · learn it in a 40-minute sprint

Pydantic AI Typed Agent

For software engineers

Pydantic AI is a Python framework for building agentic LLM applications with the same typed, validation-first feel that Pydantic brings to data models. The core object you touch is an Agent: it can carry dependency types into prompts and tools, and return a validated Pydantic output model instead of loose text.

bash

mkdir -p pydantic-ai-scout
cd pydantic-ai-scout
python -m venv .venv
source .venv/bin/activate
python -m pip install -U pip
python -m pip install pydantic-ai pytest

Scout Loop · learn it in a 40-minute sprint

Ragas Eval Loop for Scout Loops

For ML engineers

Ragas is an open-source evaluation toolkit for LLM applications. Instead of reading a few generated outputs and deciding by taste, you create a small evaluation dataset, choose metrics, run an experiment, and inspect scores and failures.

bash

mkdir ragas-scout-eval
cd ragas-scout-eval
python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install ragas langchain-openai datasets pandas

Scout Loop · learn it in a 40-minute sprint

Teachable Machine

For product-minded builders

Teachable Machine is a browser tool for training small machine learning classifiers from examples you provide. The fastest useful pass is an image project: collect a few webcam examples for two or three classes, train in the browser, test the preview, and inspect the export path.

javascript

const URL = "https://teachablemachine.withgoogle.com/models/YOUR_MODEL_ID/";
const model = await tmImage.load(URL + "model.json", URL + "metadata.json");
const prediction = await model.predict(webcam.canvas);
console.table(prediction);

Scout Loop · learn it in a 40-minute sprint

n8n: Build a Real Agent on the Canvas

For automation builders

n8n is a source-available workflow automation platform built around a visual, node-based editor. You wire triggers, integrations, logic, and AI nodes together on a canvas — each passing a stream of JSON to the next — and drop into a Code node when the visual layer runs out.

bash

npx n8n

Scout Loop · learn it in a 40-minute sprint

Claude Code: Run One Real Agentic Loop

For developers adopting AI tooling

Claude Code is an agentic coding tool that reads your codebase, edits files, runs commands, and integrates with your dev tooling. Instead of copying snippets into a chat box, it runs an agentic loop in your project: gather context, take an action, verify the result, repeat.

bash

# Install (pick one)
curl -fsSL https://claude.ai/install.sh | bash
# or: brew install --cask claude-code

mkdir scout-claude-code && cd scout-claude-code && git init

Scout Loop · learn it in a 40-minute sprint

HTTP Response Codes

For anyone working with APIs

An HTTP status code is the three-digit number on the first line of every response, and its first digit tells you the class: 1xx informational, 2xx success, 3xx redirection, 4xx client error, 5xx server error. RFC 9110 makes that grouping load-bearing — an unknown 4xx is handled like 400 — so you can switch on the first digit instead of enumerating every code.

bash

curl --version | head -1
python3 --version

Scout Loop · learn it in a 40-minute sprint

Polars: A Hands-On Scout Loop

For data engineers & analysts

Polars is a DataFrame query engine written in Rust, with first-class Python bindings. Instead of pandas' single-threaded, index-centric model, it stores data in Apache Arrow columnar memory and runs vectorized and multi-threaded by default — with both an eager API and a lazy, optimized query API.

bash

python -m venv .venv && source .venv/bin/activate
pip install polars
python -c "import polars as pl; print(pl.__version__)"

Scout Loop · learn it in a 40-minute sprint

pgvector Local RAG Store

For AI / RAG builders

pgvector is an open-source PostgreSQL extension that adds a real vector column type and similarity-search operators to a database you already know how to run. You store embeddings beside their text and metadata, and retrieval becomes an ordinary SQL query with an ORDER BY on a distance operator.

bash

docker run -d --name pgvector-lab \
  -e POSTGRES_PASSWORD=postgres \
  -p 5432:5432 --shm-size=1g \
  pgvector/pgvector:pg18-bookworm

Scout Loop · learn it in a 40-minute sprint

LangChain: Build and Evaluate a v1 Agent

For LLM app builders

LangChain is a Python and TypeScript framework for building LLM agents — programs where a model receives context, decides which tool to call, runs it, and loops until done. Since its v1.0 release the headline abstraction is a single create_agent function that wraps that loop, with durable state and streaming coming from the LangGraph runtime underneath.

bash

python -m venv .venv && source .venv/bin/activate
pip install -U langchain langchain-anthropic

What is a Scout Loop?

A bounded first pass that ends in a decision — not another open tab.

Scout

Bring a tool or idea. Informon gives you the real context and setup — no drifting through ten tabs to begin.

Try

Do one practical move with commands, code, and checkpoints — enough to feel whether it actually fits your work.

Decide

Leave with a call: Not for Me, Fits Me, Go Deeper, or shape it into a Project. Decision-quality learning.

Inside a loop

Real code you can paste and run.

Each loop teaches progressively: small blocks, a clear checkpoint, and the one bridge that turns a demo into something you could ship.

javascript

const modelURL = URL + 'model.json';
const metadataURL = URL + 'metadata.json';
const model = await tmImage.load(modelURL, metadataURL);
const prediction = await model.predict(imageElement);

Browse by what you build

Tracks for every kind of builder.

Request access →

AI & MLDev tools & CLIData & analyticsInfra & backendProduct & decisions

Stop bookmarking.Start scouting.

Sample Scout Loops. Open to everyone.

Agent Observability / Tracing Sprint

A bounded first pass that ends in a decision — not another open tab.

Real code you can paste and run.

Tracks for every kind of builder.

Stop bookmarking.
Start scouting.