Open to opportunities

Kieron Leng

@kieronleng

Message

I build machine learning pipelines for biomedical imaging and rigorous time-series forecasting.

South Africa

Message

What I'm looking for

I'm open to hybrid or full-time opportunities in 2027 where I can build reliable machine learning pipelines for biomedical data, computer vision, or data-driven forecasting.

I'm developing a U-Net-based skin-lesion segmentation pipeline for my Bioinformatics Honours research at the University of Cape Town, using the HAM10000 dataset to benchmark clinically reliable boundary delineation against expert-annotated masks.

I also built a rooftop-detection image-segmentation model with DeepLabV3+ and ResNet50, exploring how computer-vision architectures can identify cellular anomalies in biological datasets.

Alongside biomedical work, I'm building a modular Python research framework for algorithmic forex trading. I use chronological out-of-sample validation, broker-realistic cost modelling, and direct baseline comparisons to avoid misleading performance conclusions.

My projects also include genetic-algorithm optimization for a 2D physics simulation, where I achieved a class-leading score of 1500, and improving an Asteroid game through custom animations and more responsive collision physics.

Experience

Work history, roles, and key accomplishments

Bioinformatics Honours Student

University of Cape Town

Pursuing an Honours degree in Bioinformatics, focusing on machine learning pipelines for biomedical data and independent research in financial time-series forecasting.

Python Deep Learning PyTorch Computer Vision U Net Bioinformatics Algorithmic Trading Time Series Analysis