Welcome to my blog. My name is Alex Minnaar and I work in the areas of machine learning/data science/software engineering. In terms of machine learning I have experience with deep learning, natural langage processing, and recommender systems as well as ML frameworks such as TensorFlow, Keras, CoreNLP, MALLET, and Spark MLlib. In terms of software engineering I haved worked with languages such as Scala, Python, and C++ along with distributed processing frameworks like Apache Spark and low-level frameworks like CUDA C.
All of my projects (and code for this blog) can be found at my github account. Here are a few projects that I created and/or was the main contributor.
- ScaLDA: A distributed Spark implementation of the online Latent Dirichlet Allocation algorithm.
- Time Series Classification and Clustering: Python time-series k-NN classification and k-means clustering using dynamic time warping.
- Akka DistBelief: An Asynchronous implementation of the DistBelief system for training deep neural networks using the Akka actor framework.
My Open Source Contributions
Here are a few projects that I have made contributions to. I have also included links to my PRs.
- google/jax: Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more (My PRs).
- deepmind/open_spiel: OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games. (My PRs).
- deepmind/bsuite: bsuite is a collection of carefully-designed experiments that investigate core capabilities of a reinforcement learning (RL) agent. (My PRs).
- NVIDIA/PyProf: A GPU performance profiling tool for PyTorch models. (My PRs).