Build Your First Agentic Research Lab
Start with the research question, build a reproducible Python lab, account for costs, validate honestly, and turn useful engineering skills into work others can trust.
8 published lessons · Follow the sequence at your own pace.
- START HERE · FREE TO READRead lesson →
Why we started this journal: Can a trading agent actually learn?
An introduction to our research series and the evidence needed to tell learning from a lucky trading result.
- LESSON 2 · FREE TO READRead lesson →
Edition 2: Build a reproducible Python research lab
Create an isolated workspace, generate a small synthetic dataset, and leave a record another reader can reproduce.
- LESSON 3 · FREE TO READRead lesson →
Edition 3: Price changes are not profits
Build a transparent paper ledger, distinguish gross from net outcomes, and see why costs can overturn a promising-looking result.
- LESSON 4 · FREE TO READRead lesson →
Edition 4: Freeze the rule before you judge it
Separate rule selection from later evaluation, keep a baseline, and publish an honest result when a challenger fails.
- LESSON 5 · FREE TO READRead lesson →
Edition 5: Give the agent a boundary, not a blank check
Separate generated suggestions from validated decisions, reject unsupported evidence, and make an agent workflow inspectable.
- LESSON 6 · FREE TO READRead lesson →
Edition 6: Make automation observable and retries deliberate
Record job states, distinguish retryable failure from an uncertain outcome, and prove that a repeated run does not duplicate a completed local result.
- LESSON 7 · FREE TO READRead lesson →
Edition 8: Earn from research skills, not return promises
Package a reproducible lab, workshop, or software service around a clear customer problem and an honest scope.
- LESSON 8 · FREE TO READRead lesson →
Edition 7: Publish a research report someone can reproduce
Turn a small experiment into an inspectable report with a data fingerprint, assumptions, checks, and limitations.
Learn at your own pace.
Each lesson is an article you can revisit. Use the previous and next links to stay oriented. Member editions use the publication's existing subscription access.
Follow the publication for future editions ↗