Human Feedback for AI EvaluationIn progress
A framework for evaluating AI-generated outputs, covering annotation guidelines, inter-rater agreement, quality scoring and failure-mode analysis, with reproducible Python workflows.
Research program leader · Strategy & operations · Quantitative research
I study how animals signal and how complex systems respond to social and environmental pressure. My PhD research asked how elaborate sexual coloration is structured and what it does in an endangered primate, the drill, combining field observation in Nigeria, calibrated image analysis and the first chemical characterization of drill scent-gland secretions. I also design and run the programs behind the science: international collaborations, field operations, budgets, permits and the reproducible analysis pipelines that turn raw data into decisions.
I spent more than a decade in research leadership. That work ranged from field data collection in East and Southern Africa to designing and leading a multi-year international research program across Nigeria and the United States. Now I bring that experience to strategy and operations, where I turn data-heavy, ambiguous problems into structured programs with clear metrics, reproducible workflows and aligned stakeholders.
A framework for evaluating AI-generated outputs, covering annotation guidelines, inter-rater agreement, quality scoring and failure-mode analysis, with reproducible Python workflows.
An end-to-end pipeline that takes raw field images through image processing, feature extraction and statistical validation to publication-ready results, built for hundreds of field-collected images.
A reproducible reporting system that generates figures, tables and manuscripts directly from raw data, which cuts manual reporting time and copy-paste errors.
A multi-year, multi-institutional research program spanning field sites, laboratories and partner organizations in two countries, including its budget, logistics, compliance and a distributed team.
I study how animals signal to one another and how complex signals evolve. My dissertation focused on drills (Mandrillus leucophaeus), an endangered West African primate with some of the most elaborate sexual coloration of any mammal. It asked how that coloration is structured, how stable it is, and what it does in competition and mating. I also made the first characterization of the chemical compounds in drill scent-gland secretions. Alongside the science, I build the field programs and analysis pipelines that make the data trustworthy.
Male drills show striking red, blue and purple coloration and are far larger than females. Yet little was known about how this coloration is organized, how it changes over time, or how it relates to social rank and reproduction. I asked three things: what the signal looks like (its form and structure), what it does between males (whether it signals status and helps males size up rivals), and what it does between the sexes (whether it predicts mating success and female interest).
Mixed-effects models, longitudinal analysis, experimental design, calibrated image analysis and visual modeling, behavioral sampling, dominance-rank estimation, and reproducible workflows in R and Python.
See the full list on Google Scholar.