Curriculum Vitae
Applied AI, data science, AI engineering and research. Nine years building and operating production machine learning and data systems across insurance, telecommunications, applied research and higher education, and increasingly turning those lessons into reusable open source infrastructure.
What the work delivered
Production systems that survive noisy data, organisational constraints and real people using them.
R1B
Approx. Annual Savings
Vodacom Smart Generator Optimisation
100M+
Events a Day
Ubunye Engine, powering 10+ products at a bank
R2M+
Annual Subsidy Impact
Wits recommendation system
300K+
Audience Reached
FabAcademic Unfiltered
Absa
Mar 2024 - Present
Lead Data Scientist, Absa Group
Ubunye Engine at 100M+ events a day for 10+ products · telematics from months to under a day
Vodacom
Nov 2021 - Mar 2024
Senior Data Scientist, Vodacom
≈R1B annual operational savings · tens of millions of events a day across 15,000+ sites · Vodacom Star Award 2022
IBM
Apr 2020 - Nov 2021
Machine Learning Research Scientist, IBM Research
Models into a petabyte-scale geospatial platform · provincial COVID-19 planning dashboard · NeurIPS 2020 CCAI workshop paper
The route so far
From a stock exchange in 2017 to a PhD in 2027. Tap a stop to see the work done there.
PhD in Computer Science
University of the Witwatersrand
Commencing in 2027; the proposal is in preparation and not yet registered. The proposed research explores physics-informed self-supervised learning for SAR-based flood extent mapping, with applications to data-scarce climate and insurance-risk settings. It follows on from the MSc work on echo state networks for level set segmentation, and from the geospatial and climate risk systems built at IBM Research and in insurance.
Proposed research: physics-informed self-supervised learning for SAR flood mapping
Lead Data Scientist
Absa Group
Took data science from infrastructure not built for big data to a modern platform: led the modernisation strategy and architected the data engineering capability from scratch, bringing Databricks into the bank with CI/CD and integration with legacy systems, as part of the move from on premises to cloud (AWS and Databricks). That capability runs on Ubunye Engine, which processes 100M+ events a day and is the engineering layer for 10+ products. Modernised telematics, about 7 to 10 million trip records a day, from months to under a day, and the flood forecasting used by underwriting. Leading the hyper personalisation strategy with customer lifetime value and segmentation models. Lead a team of three data scientists and introduced AI and model governance aligned to the bank’s standards.
Ubunye Engine at 100M+ events a day for 10+ products · telematics from months to under a day
Read the storySenior Data Scientist
Vodacom
Led a team of 10 data scientists and ML engineers building real-time analytics, optimisation and decision-intelligence systems for national telecommunications infrastructure, combining machine learning, mathematical optimisation, streaming data and software engineering on problems including generator optimisation, traffic forecasting, infrastructure planning, anomaly detection, resource allocation and site prioritisation. Led the Smart Generator Optimisation work across more than 15,000 sites, contributing approximately R1 billion in annual operational savings through better allocation and use of mobile power infrastructure. Built and designed streaming systems processing tens of millions of daily telemetry and alarm events using Kafka, PyFlink, PySpark and Kubernetes. Established reusable engineering practices and technical standards across the team, and mentored practitioners while working closely with network, engineering, operations and business stakeholders. Received the Vodacom Star Award in 2022 for engineering contribution and impact.
≈R1B annual operational savings · tens of millions of events a day across 15,000+ sites · Vodacom Star Award 2022
Read the storyMachine Learning Research Scientist
IBM Research
Conducted applied machine-learning research in climate, environmental intelligence, remote sensing and geospatial analytics, combining scientific experimentation with production-oriented engineering. Worked with large geospatial, satellite and environmental datasets in distributed environments using TensorFlow and related tooling, developing forecasting and predictive models that could operate beyond single-machine research workflows. Deployed climate-forecasting models into IBM PAIRS Geoscope, connecting research outputs to a petabyte-scale geospatial-temporal platform used for environmental and enterprise analytics. Contributed to the Gauteng COVID-19 risk-index and prediction dashboard with IBM Research Africa, Wits University and the GCRO, supporting hotspot identification and healthcare-resource planning for the Gauteng Provincial Department of Health. Co-authored research with international scientists and engineers, contributing across experimentation, model development, evaluation, data pipelines and operationalisation.
Models into a petabyte-scale geospatial platform · provincial COVID-19 planning dashboard · NeurIPS 2020 CCAI workshop paper
Read the storyData Scientist
Business Intelligence Services - University of the Witwatersrand
Built analytics, reporting and machine-learning systems for university-wide institutional planning, student success and decision support using Python, SQL, Power BI and statistical modelling. Developed student recommendation and progression analytics to help faculties and institutional teams see where students needed support and make better academic planning decisions. Built a clustering-based recommendation system for the Faculty of Humanities, linking analytical work to student progression and government subsidy outcomes, with more than R2 million in annual subsidy impact attributed to the initiative. Facilitated analytics and machine-learning workshops for staff and students, helping institutional researchers and analysts adopt practical modelling in their own work.
R2M+ annual subsidy impact attributed to the initiative · analytics and ML workshops for staff and students
Read the storyMSc in Computer Science (Distinction)
University of the Witwatersrand
Thesis: Learning Level Set Method by Echo State Network for Image Segmentation. Proposed a novel spatiotemporal deep-learning formulation of variational level set segmentation, benchmarking five recurrent and convolutional architectures.
Distinction · Echo State Networks for image segmentation
Data Scientist & Software Engineer
Council for Scientific and Industrial Research (CSIR)
Built Django-based predictive analytics and decision-support systems serving 17 municipalities, including the City of Cape Town and 16 across Gauteng, giving public-sector stakeholders operational visibility and real-time access to analytics they could plan against. Applied machine learning and data engineering to surface operational bottlenecks and improve service-delivery planning, working in multidisciplinary teams spanning software engineering, analytics and public-sector innovation.
Recognised by Mail & Guardian, CSIR & DST for innovation
Read the storySystem Analyst & Support
ZAR X
Provided systems analysis and operational support for a newly launched stock exchange serving 13,000+ clients, keeping the trading platform stable through go-live and early operations as requirements changed week to week. An early foundation in financial systems, integration, documentation and operational reliability.
Operational support for a newly launched stock exchange serving 13,000+ clients
BSc Honours in Computer Science
University of the Witwatersrand
Project: Wildfire Estimation Using Kernel Density Estimators.
Wildfire estimation with kernel density estimators
BSc in Computational & Applied Mathematics and Astronomy
University of the Witwatersrand
Mathematical modelling, astrophysics, simulation, and numerical methods.
Foundations in mathematics, astrophysics & simulation
What people say
From colleagues
“Thabang was an integral member of my team for several years, during which we tackled numerous challenging data science problems together. His dedication and passion for solving complex problems are truly commendable. His technical knowledge in data science is impressive, with hands-on experience in Gaussian process models, graph algorithms, probabilistic graphical models, and traditional machine learning algorithms.”
Jaco du Toit, Ph.D.
AI/ML & Data Team Lead, Vodacom
Impact“It was evident from the start that Thabang is an exceptional and hard-working researcher driven by achieving the goals set before him. He always found a way to get a task done — he took initiative, made a decision, and executed on it with speed and to the best of his ability. It was an absolute pleasure to have worked with him.”
Etienne Vos, Ph.D.
Research Scientist Manager, IBM
Teaching“As a data scientist, he played an instrumental role in enhancing our data and prediction pipelines. He identified new sources of predictability and tested multiple hyper-parameter search strategies, which ultimately boosted the performance of our models. His attention to detail, willingness to collaborate, and strong communication abilities made him an invaluable teammate.”
Akram Zaytar, Ph.D.
Senior Research Scientist, Microsoft | GeoAI
Engineering