Angus Mackenzie


About Me

I’m Angus, an SDE working at AWS. I am from Johannesburg, currently residing in Cape Town, and I enjoy meeting people and helping others. I am always keen to gain experience, and help solve problems.

Some things I currently have experience with are: Java, Python, Machine Learning, React, and Eating Pizza


Bayesian Approximation of Effective Machine Learning

I completed this project as part of my Adaptive Computation and Machine Learning course at Wits. We were tasked with using some of the course’s content to create something useful for us and other students in the class. I decided to use Recurrent Neural Networks (RNNs) as a means to generate titles for research papers, or any other writing. I compared two RNNs to a Long Short Term Memory (LSTM) model, and was able to create some fun and interesting titles as well as learn more about RNN’s and different neural network architectures. Some of my favourite titles generated are as follows:

  • Function Classification Using Machine Learning Techniques
  • Bayesian Approximation of Effective Machine Learning
  • Computer Multi-agent Boltzmann Machine Learning
  • New Machine Learning for Astrophysics

Ekans ID

This project was completed during my Computer Vision course at Wits. We had to identify as many different snakes from the Snake Identification challenge dataset as possible. We could use any technique we wanted, as long as we were able to achieve reproducible, unique and interesting results. We achieved some really great results using EfficientNet, and explore some really cool visualisation techniques like Class Specific Image Generation, and Occlusion Sensitivity.

We created a report, which is available on the repository linked above.




May 2020 - Current

UX Platform Team

I am currently a SDE at AWS, working on developing internal tools to aid and assist software development teams working on the consoles.



January 2020 - April 2020

Allowlist me pls.

Worked on baking some features into an internal allowlisting tool for my team that allowed users to easily add an audit trail for the changes they were making and propagating. Worked with React, typescript and java. My first time working on complex frontend systems, and I really enjoyed getting a grip on how all the different aspects came together.



November - December 2018

Whats all the FaaS about?

Used Kubeless as a framework to create a Function as a Service (FaaS) metric gathering tool. Used Docker to modify the kubeless-python runtime. Wrote functions within that runtime. Used Serverless to launch functions, and attempted to create a react web-app to deploy functions.



July 2018

JMXtra cool

I worked with Java Management Extensions (JMX) in order to capture metrics for processes. Wrote a proxy-creator and instrumented it with JMX to ensure that whenever anyone adds a new resource; metrics are captured for that resource automatically. Used ElasticStack to capture the JMX metrics, and then visualized them with Kibana. Worked briefly with Ubuntu Juju and Xen

MWR Infosecurity


June 2018

Established in 2003, MWR is an independent cyber security consultancy

Tested a web application for vulnerabilities. Made a report of the vulnerabilities I found. Developed proof-of-concepts for an inference attack and a login backdoor in python and php. Presented a summary of my report and demonstrated my proof-of-concepts.



July 2017 - August 2017

Established in 2001, RSAWEB is a Internet Service Provider

Implemented a Jira platform for their teams. Integrated it with their previous system. Coded in PHP, utilized Jenkins and Chef, worked with MySql, worked in a team.



Honours in Computer Science


University of Witswatersrand

I researched Deep Similarity Learning for Unique Animal Identification, with the aim of identifying unique individuals within a population of Leopards during my honours. I created a dataset of leopards with the help of guides from the Sabi Sands Wildtuin. It was a really interesting project, and gave me a good handle on current animal identification practices as well as more holistic understanding of machine learning.

I really enjoyed Wits, and had a lot of fun with their very practical courses.


BSc in Business Computing & Computer Science

2015 - 2018

University of Cape Town

I double majored in Computer Science and Information Systems. Being exposed to so many different people and experiences really made my entire stay at UCT amazing.

Whilst studying at UCT I was part of the Social Tennis club, Developer Society as well as the Crypto and AI society. I also tutored Computer Science privately - with my tutees seeing a consistent rise in marks and all of them passing their courses.

Some More About Me

I grew up in Johannesburg, and for as long as I can remember I have been tinkering with anything I can get my hands on, especially with tech. Information Technology in high school really piqued my interest and since then I have been entirely enthralled by computers, how they work and programming them.

When I am not working I like to socialize, work on cars, or spend time outdoors. I really like fly fishing, camping and hiking.

This website was created using the Modern Resume Theme website template made by James Grant