I am a PhD researcher in the Cyber Physical Intelligence Lab of the Engineering Systems and Services department at the Technical University of Delft. I work under the supervision of Dr. Aaron Ding and am promoted by Prof. dr. ir. Marijn Janssen. I do interdisciplinary research at the intersection of applied machine learning, edge computing and embedded systems, with a focus on trustworthy machine learning systems in the Internet of Things.
I view machine learning systems as powerful technology components that can enable large, engineered systems to be flexible, personalised and adaptive. As such, they must be trustworthy and behave reliably in complex and safety-critical applications. I seek to understand their characteristics and limitations by building testing frameworks to characterise trustworthiness aspects, like fairness. At present I am working on model compression and domain adaptation techniques for on-device inference of audio models, with the goal of building deeply personal and completely private voice assistants. I am interested in voice-based applications for smart homes, health and elderly care, gender-based violence and large scale behavioural analytics in cities.
I received my full research masters in computer science (artificial intelligence) with distinction from the University of Cape Town under the supervision of Assoc. Prof. Deshen Moodley and Prof. Tommie Meyer. Outside of my academic research I am the co-founder of Engineers Without Borders South Africa, a non-profit organisation with the mission to empower engineers to empower communities. I am a fellow of TechWomen (Silicon Valley, 2018) and Data Science for Social Good (London, 2019), and am a project lead in Mozilla’s Trustworthy AI Working Group. I am always open to conversations about AI and consciousness, large scale infrastructure development and transformation.
|Jun 20, 2022||I am at ACM FAccT to present our paper on Bias in Automated Speaker Recognition.|
|Jun 15, 2022||Our paper with Design Guidelines for Inclusive Speaker Verification Evaluation Datasets has been accepted to Interspeech 2022.|
|Jan 27, 2022||My new project FairEVA has been selected for the inaugural cohort of the Mozilla Technology Fund!|
|Jan 21, 2022||Read our pre-print on reliability bias in tiny ML.|
|Nov 17, 2021||At AIChallengeIoT 2021 at SenSys to talk about pre-processing parameters in tiny audio.|
|Jun 19, 2021||Off to Cambridge to start my summer internship at Nokia Bell Labs!|
|May 25, 2021||Superb panel discussion at Rethink AI with Mozilla and Women in AI Ethics. Here’s a write up.|
|Mar 8, 2021||MozFest 2021 is finally here! Join us at the AI In Real Life Space.|