Four weeks. Twenty-eight functions. One capstone. What's the pattern that ties everything together?
Agent as a function. Every week was the same idea — one agent call, one new thing added. system_prompt, result_type, batch, search, tools. Each week extended the previous one.
That's the track's architecture. Week 1 built the primitive. Week 2 added structure. Week 3 added iteration. Week 4 added search and decision-making. Each layer is a function you can use independently or chain.
I know the pattern for verifying AI output: validate a sample, check fields against sources, disclose the methodology. The capstone is a tool, not a replacement for research judgment.
Six questions. Four from Week 4, two reviews. Then your research assistant is complete.
I built something I'll actually use in my thesis. That's a better outcome than I expected from a 15-minutes-a-day track.
The track was designed for that outcome. Let's confirm the patterns hold.
Create a free account to get started. Paid plans unlock all tracks.
Four weeks. Twenty-eight functions. One capstone. What's the pattern that ties everything together?
Agent as a function. Every week was the same idea — one agent call, one new thing added. system_prompt, result_type, batch, search, tools. Each week extended the previous one.
That's the track's architecture. Week 1 built the primitive. Week 2 added structure. Week 3 added iteration. Week 4 added search and decision-making. Each layer is a function you can use independently or chain.
I know the pattern for verifying AI output: validate a sample, check fields against sources, disclose the methodology. The capstone is a tool, not a replacement for research judgment.
Six questions. Four from Week 4, two reviews. Then your research assistant is complete.
I built something I'll actually use in my thesis. That's a better outcome than I expected from a 15-minutes-a-day track.
The track was designed for that outcome. Let's confirm the patterns hold.