LAB/01
AI experiments
Small, focused tests of model behaviour, prompting and tool use.
LLMTool use
Labs
Experiments, benchmarks and ideas that are useful even when they do not become full projects.
LAB/02
.NET benchmarks
Measuring runtime, library and pattern choices instead of guessing.
BenchmarkDotNetPerformance
LAB/03
RAG experiments
Chunking, ranking and grounding strategies compared side by side.
RetrievalEvaluation
LAB/04
Coding-agent evaluation
How well coding agents handle real-world .NET codebases and tasks.
AgentsBenchmarks
LAB/05
Architecture prototypes
Throwaway systems built to test a design idea before committing to it.
DesignPrototypes
LAB/06
Developer tooling
Utilities that remove friction from everyday engineering work.
CLIAutomation
No experiments published yet. Results will be linked from each area as they are written up.
How experiments are run
STEP 1
Question
Start from a concrete engineering question, not a technology.
STEP 2
Measure
Prefer reproducible numbers and published harnesses over impressions.
STEP 3
Write down
Record what was learned, including the experiments that didn't work.