Topics
The subjects the work is about, across projects, research, publications, writing and talks. A subject gets its own page once there is enough work behind it to be worth reading; the rest are listed so the full range is visible, and each links to wherever that work lives.
Topic pages
- AI education and practical AI literacyPractical AI education from South Africa: FabAcademic Unfiltered sessions, Kasilam township capability building, and university ML workshops.17 items
- AI agentsBuilding with AI agents in practice: agentic systems, retrieval, tools, security, reliability and CI, plus public sessions for non coders.10 items
- PythonPython as an engineering language: open source libraries for Spark pipelines and Bayesian filtering, and how to package a library properly.8 items
- Software engineeringSoftware engineering for AI systems: agents in the development workflow, reliability, security, CI, packaging, and open source code to inspect.6 items
- Recurrent neural networksRecurrent neural networks across three research lines: hearing loss in mine workers, seasonal climate forecasting and echo state networks.6 items
- Climate risk and geospatial machine learningClimate risk and geospatial ML: seasonal forecasting research, models in a petabyte scale platform, and flood risk across 230,000+ properties.6 items
- Data scienceApplied data science in insurance, telecoms, research and higher education, and writing on building the capability around the models.4 items
- Data engineeringData engineering behind production analytics: streaming telemetry, Spark and Databricks pipelines, and the open source Ubunye Engine.4 items
- MLOps and production machine learningProduction machine learning from practice: governed ML on Databricks, real time systems at telecoms scale, and the open source Ubunye Engine.3 items
- Computer visionComputer vision research: learning level set image segmentation with echo state networks and recurrent models, plus an object following robot.3 items
- Open sourceOpen source Python from production problems: Ubunye Engine for portable Spark pipelines, TFiltersPy for Bayesian filtering, and how to package.3 items
- Apache SparkApache Spark in production and open source: telecoms telemetry, insurance workloads on Databricks, and portable pipelines with Ubunye Engine.3 items
More fields
More methods
- Retrieval augmented generation: AI education: FabAcademic Unfiltered and Thabang AI, The Practical Roadmap to Building With AI Agents, Part 3: Giving an Agent Knowledge
- Reservoir computing: Echo state networks for level set image segmentation, Learning Level Set Method by Echo State Network for Image Segmentation
- Echo state networks: Echo state networks for level set image segmentation, Learning Level Set Method by Echo State Network for Image Segmentation
- Image segmentation: Echo state networks for level set image segmentation, Learning Level Set Method by Echo State Network for Image Segmentation
- Level set method: Echo state networks for level set image segmentation, Learning Level Set Method by Echo State Network for Image Segmentation
- Deep learning: Climate forecasting and geospatial ML at IBM Research, Echo state networks for level set image segmentation, Machine learning for seasonal climate forecasting
- Bayesian filtering: TFiltersPy: Kalman and particle filters for Python
- State estimation: TFiltersPy: Kalman and particle filters for Python
- Kalman filtering: TFiltersPy: Kalman and particle filters for Python
- Ensemble Kalman filter: TFiltersPy: Kalman and particle filters for Python
- Particle filtering: TFiltersPy: Kalman and particle filters for Python
- Time series: TFiltersPy: Kalman and particle filters for Python, Machine learning for seasonal climate forecasting
- Mathematical optimisation: Generator optimisation and streaming at Vodacom
- Recommender systems: Student success analytics and recommendations at Wits
- Stream processing: Generator optimisation and streaming at Vodacom
- Extract, transform, load: Ubunye Engine: portable Spark pipelines for data and ML
- Seasonal forecasting: Climate forecasting and geospatial ML at IBM Research, Machine learning for seasonal climate forecasting, ML-based Probabilistic Prediction of 2m Temperature and Total Precipitation
- Reproducibility: Ubunye Engine: portable Spark pipelines for data and ML
More domains and problems
- Technical leadership: Generator optimisation and streaming at Vodacom, Insurance data science: telematics, flood risk and MLOps
- Geospatial machine learning: Climate forecasting and geospatial ML at IBM Research, Insurance data science: telematics, flood risk and MLOps, Machine learning for seasonal climate forecasting
- Remote sensing: Climate forecasting and geospatial ML at IBM Research, Proposed: self supervised flood mapping from radar
- Synthetic aperture radar: Proposed: self supervised flood mapping from radar
- Telematics: Insurance data science: telematics, flood risk and MLOps, TFiltersPy: Kalman and particle filters for Python
- Decision support systems: Decision support systems for municipalities at the CSIR, Generator optimisation and streaming at Vodacom, Student success analytics and recommendations at Wits
- Noise induced hearing loss: Machine learning for hearing loss risk in mine workers, Mine workers threshold shift estimation via optimization algorithms for deep recurrent neural networks, Noise level policy advising system for mine workers
- Occupational health and safety: Machine learning for hearing loss risk in mine workers, Mine workers threshold shift estimation via optimization algorithms for deep recurrent neural networks, Noise level policy advising system for mine workers
- Telecommunications: Generator optimisation and streaming at Vodacom
- Insurance: Insurance data science: telematics, flood risk and MLOps
- Public sector: Decision support systems for municipalities at the CSIR
- Capacity building: Kasilam: teaching township youth to build with AI
- Robotics: Follow me Turtle Bot Project
More technologies
- Databricks: Insurance data science: telematics, flood risk and MLOps, Ubunye Engine: portable Spark pipelines for data and ML
- Kubernetes: Generator optimisation and streaming at Vodacom, Ubunye Engine: portable Spark pipelines for data and ML
- Docker: Generator optimisation and streaming at Vodacom, Ubunye Engine: portable Spark pipelines for data and ML
- Dask: TFiltersPy: Kalman and particle filters for Python
- Apache Kafka: Generator optimisation and streaming at Vodacom
- Apache Flink: Generator optimisation and streaming at Vodacom
- MLflow: Insurance data science: telematics, flood risk and MLOps
- TensorFlow: Climate forecasting and geospatial ML at IBM Research
- Django: Decision support systems for municipalities at the CSIR
- NumPy: TFiltersPy: Kalman and particle filters for Python