The boundary is where modernization risk lives
A controlled C++/.NET interop study that separates boundary hardening from compiler settings, so a hostile-input comparison and a compiler comparison each isolate exactly one variable.
Engineering notes
Engineering decisions,
examined.
Short analyses of design alternatives, evaluation findings and operational trade-offs from my independent engineering work. Each note links to the underlying implementation and its stated limits.
Diagnosing where a retrieval pipeline loses evidence
An offline evaluation lab that attributes evidence loss to a specific pipeline stage — chunking, retrieval or ranking — using a synthetic corpus with known evidence spans and a predict-then-measure discipline.
Detecting AI answer-quality regressions that don't throw errors
A simulation of the gap between request-level monitoring and answer-quality monitoring, comparing aggregate and topic-sliced drift detectors on deterministic, non-semantic features under a modeled operating budget.