AI. Data. Social Value.
I Bridge the Gap Between AI’s Potential and Jagged Every-day Life.
I lead teams that transition research-grade models to AI solutions with systemic impact. With a foundation in high-fidelity data collection and rigorous evaluation, I look beyond technical silos to build solutions that are fit-for-use, resilient, and socially grounded.
Designing responsibly since 2010.
Currently Senior Research Scientist at Sony AI
The greatest challenge of our time is to align the infinite possibilities of technology with a future where life and humans flourish.
My superpower is to mobilize people towards a common project purpose, and to deliver on that purpose effectively and timeously. With a PhD (cum laude) from TU Delft in the Netherlands, a background in computer science and foundations in mechanical engineering, I pursue practical and purposeful technology development for public value. For the past decade my work has focused on fit-for-purpose AI evaluation and design choices for responsible AI development.
Strategic Pillars
Systems Leadership & Orchestration
I am a self-starter and take a high degree of initiative to bring diverse stakeholders together to realize complex projects. I excel at identifying strategic leverage points for value creation and coordinate team members and partners to deliver outcomes that are trustworthy and impactful.
Data Strategy & Stewardship
Quality AI products need quality data. From building technical data portals, to developing specifications for data collection and annotation, to quality control and data governance, I have extensive experience in realizing technical and stakeholder inputs for high-integrity data ecosystems.
AI Benchmarks & Fit-for-Use Evaluation
AI that’s technically correct but practically useless? The only way to find out is to evaluate. I approach evaluation holistically, working from first principles to consider metrics and scenarios that matter to end-users. This means trading single performance metrics for multi-objective considerations.
Featured Projects
Computer Vision for the Global Majority
Collecting the world’s first ethically sourced, globally diverse computer vision dataset of people.
AI Risk & Reliability Benchmarking
Systematic grading and scoring for LLM safety assessments for the ML Commons AILuminate Benchmark.
Design Patterns for AI Bias Evaluation
Establishing best practices for bias detection and mitigation in AI for practitioners and developers (PhD thesis).
Big Data for Household Energy Insights
Published South Africa’s first large-scale, longitudinal dataset of residential electricity demand from 1994 to 2014.
Data for Municipal Infrastructure
Leveraging data to support municipal infrastructure policy in South Africa, shifting 15% of national infrastructure budget.
People-Centric Engineering Design
Transforming engineering education through locally embedded design challenges and international collaboration.
