Rails, AI & Product by Viktor

Sign in Subscribe

Diagram: one expensive planning phase feeds a repeated cheap-model build phase of code, tests, and bug fixes.

ai-augmented-development

Plan With the Frontier Model, Build With the Cheap One

Claude Fable 5 ran a nonprofit's club-management system through Spec Kit's full round trip, twice, unattended, inside a free access window that would otherwise cost about $85. A cheaper model, picked for being the best fit for the language, builds it this weekend.

Five-node flow showing a Task call passing through model routing, on-demand loading, and reference passing to a gate verdict.

ai-augmented-development

Three Levers That Shrink the Token Bill of a Claude Agent Swarm

Token cost in a Claude Code swarm is a budgeting problem, not an intelligence problem. Three levers, model routing, progressive disclosure, and context packets, keep long migrations affordable and fast.

Loop diagram showing failure, categorize, add regex pattern, validator gate, future runs skip, cost drops

ai-augmented-development

Claude Skills Made My Migration Swarm Stop Re-Arguing Itself

A multi-agent RSpec to Minitest swarm kept relearning the same conventions. Moving validation from prompts to regex-enforced gates turned every failure into a permanent fix.

Diagram showing 8 scattered tools consolidating into one dashboard

ai-augmented-development

Engineering Knowledge Compounds When Claude Code Is the Runtime

You make a decision in February. By April, three people re-debate it because the context lived in a Slack thread. I built a Knowledge OS on Claude Code where capturing knowledge is a side effect of work. After 2.5 months: 188 artifacts, 2,500+ docs in semantic search, decision chains that compound.

Flowchart showing 4 error-handling layers from generation to human intervention

ai-augmented-development

Layered Loops: Why 764 AI Agent Sessions Still Needed 21 Human Interventions

764 Claude sessions across ~259 unique files over 16 working days. The orchestrator was a bash while loop. The real engineering was in layering error handling: a generation loop, a fix loop, a cleanup orchestrator, and a human catching what all three missed.

Flowchart showing a 6-gate pipeline from Analysis through Report with color-coded stages

ai-augmented-development

From 6 Minutes to 66 Seconds: Migrating 9,835 Model Tests to Minitest with an AI Agent Swarm

RSpec model specs: 6 minutes for 3,780 examples. Minitest model tests: 66 seconds for 9,835 tests. Both parallel. I built a 6-gate multi-agent pipeline in Claude Code to make the migration work.

Flowchart showing four system phases: Sequential, Parallel, Validate & Optimize, and Test. An Analyst node splits into parallel Test Writers, merging back for validation and testing.

ai-augmented-development

Building Mathematical Testing Agents for RSpec

I built a multi-agent system that applies mathematical testing techniques to 60+ RSpec models at scale. The architecture worked. The tests were rigorous. Then the results revealed the bottleneck was never test quality.

Diagram comparing bloated factory cascade with lean refactored factories using traits

TestProf Cut Our Slowest Specs by 95%, But the Suite Still Took 30 Minutes

We used TestProf to diagnose a Rails monolith where factories consumed 95% of test time. Factory refactoring and let_it_be delivered 50-95% speedups on the worst specs. The full suite improved by 14%. The gap between those two numbers explains why manual optimization hits a ceiling.