M.Sc. Graduate, Google DeepMind Scholar
Stellenbosch University & AIMS South Africa
Actively seeking PhD opportunities
salakojoshua1234[at]gmail[dot]com
Hi 👋🏿 I recently completed my M.Sc. in Mathematical Sciences (with specialization in AI for Science) at AIMS South Africa and Stellenbosch University, graduating with distinction. I am grateful to have been supervised by Dr Lorène Jeantet. I work on machine learning for ecology, mostly computer vision for making sense of data from camera traps.
For my master's, I taught models to spot endangered Western Leopard Toads across 1.2 million images captured by camera traps, then used those detections to test whether tunnels beneath roads actually help the toads avoid roadkill. A very interesting project, and perhaps the most consequential game of “find the toad” I will ever play. I am now working with the Machine Learning for Ecology group at AIMS and looking for PhD opportunities. I am very open to conversations!
I am also curious about agentic systems. Alongside research, I am an AI/ML Engineer at Rio Eigen (industry experience pays too, quite literally). Away from models, you will usually find me behind a camera or somewhere inside a good playlist.
Third Workshop on Computer Vision for Ecology (CV4E), ECCV 2026.
Combines contrast enhancement, domain-specific transfer learning, active learning, and spatial postprocessing to detect endangered Western Leopard Toads in camera-trap imagery and monitor wildlife tunnel use.
AGRIS on-line Papers in Economics and Informatics, 16(2), 121–131, 2024.
Tests modern object detectors on catfish-pond imagery, then pairs the strongest model with tracking to automate fish monitoring for aquaculture managers.
Our paper on annotation-efficient detection of endangered Western Leopard Toads was accepted to the CV4Ecology workshop at ECCV 2026. I attended the workshop in Malmö and presented the work as a poster. [LinkedIn post]
Featured in Cape {town} Etc and Nature Connect for our work using computer vision to provide automated evidence that endangered Western Leopard Toads are successfully using wildlife underpasses.
Attended the Deep Learning Indaba 2026, at Pan-Atlantic University in Lagos. I presented Toads, Tunnels, and Transformers: Deep Learning for Detection and Identification of Western Leopard Toads, a poster based on my master's thesis, and it was selected among the top 50 prize-winning posters.
Graduated with distinction from AIMS South Africa with a master's in Mathematical Sciences, specializing in AI for Science.
Sep 2025 – Jul 2026
Stellenbosch University & AIMS South Africa
Google DeepMind Scholar · AI for Science · Distinction
Thesis: Automatic Detection and Identification of Western Leopard Toads.
Aug 2019 – Sep 2024
Federal University of Agriculture, Abeokuta, Nigeria
CGPA: 4.40/5.00
2025
Two awards recognising innovation and impact.
2025
Full scholarship for the AI for Science programme at AIMS South Africa.
2025
Scholarship in Artificial Intelligence and Data Science.