Independent verification of AI-produced analytical work: document review pipelines where no model audits its own family's output, with findings that cite their sources and survive being questioned. Built once as Project Shimmer; adaptable to a domain and a corpus.
Evaluation design for AI systems and for research claims: benchmarks with ground truth built by construction, de-primed test setups, and causal-integrity audits that ask whether a stated result survives its own unreported robustness checks.
The methodology and measurement side of proposals and programmes: monitoring and evaluation design, defining outcomes that can actually be detected, survey instruments, and the analysis plan that makes a funder's evidence requirement satisfiable. Current work in this line is the measurement angle of the Values for Cohesion project at the University of Birmingham, funded by the UK Ministry of Housing, Communities and Local Government.
Remote from İzmir, English and Turkish, invoicing as an independent contractor. Scoped engagements preferred: a defined corpus, a defined question, a defined deliverable. Write to me with two sentences on the problem, and I will tell you whether it is my kind of work.