Daniel Mahandana
Johannesburg, South Africa
I'm a Data Engineering Intern at Wits Enterprise and an Information Management graduate from the University of Johannesburg.
I design and operate resilient data pipelines, automated schema validation, and reliable storage systems that ensure clean data flows to downstream analytics and machine learning models.
Outside of production pipelines, I build experimental research software from first principles — studying self-play reinforcement learning with Monte Carlo Tree Search, latent representation geometry, and Bayesian active capability inference.
Data Infrastructure & Pipelines
Building robust ETL workflows, ingestion frameworks for complex public datasets, schema diffing, and vector indexers that withstand production failures.
Machine Learning & Search Systems
Developing AlphaZero-style neural network search engines, MCTS exploration dynamics, and probabilistic capability evaluation under resource constraints.