Know More
About Me

Hey guys Julien here! Machine Learning Engineer. I love everything Data Science, and Software Engineering related works. In Software Engineering, I do mostly back-end work jobs. In Machine Learning, I build models, test them and make them production ready. My passion is to see my code running to solve a real world problem.

I am also a Machine Learning researcher. Besides using libraries like TF, Scikit-Learn and the others, I also have a duty of improving and discovering new learning structures.

View my CV

Experience

Machine Learning Engineer
Capitec Bank - Business Bank - Client Experience Delivery : 2022-Present
Currently working on building new data driven software for Business Bank.
Machine Learning Engineer
Entelect - Standard Bank Corporate and Investment Banking - Digital Insights: 2020-2022
Lead a small team of Data, Software Engineers and Data Scientists to build FX Tips. A data driven tool tips corporate clients to buy/sell Forex products. The project is a fully in production and has brought in some revenue (around $300000) that made it profitable in its first 6 months of launch. I wrote the very first lines of code to the project and orchestrated the different companents and pipelines needed for predicting propensity to buy or sell a forex product. Overall the Python project was surrounded by CI/CD, cloud, Tensorflow, Git Control-M.
Software Engineer Engineer
Entelect - Standard Bank Corporate and Investment Banking - Risk: 2019-2020
In a team of 3 people working on automating the generation of forex reports for Standard Bank Risk teams internal used. The initial report were done using a specialized licenced software that was expensive long term. The aim was to build in in-house version of that report generation. I was responsible to write the main backend codebase to hold all the logic. That would involve building the structure of the backend server and writing the different API endpoints for each use case. This Docker-based API project is in production and performs similar to the initial software bought for the purpose.
PhD Candidate
University of the Witwatersrand: 2019-Present
Investigating the feasibility and perfomance of using Binary Neural Networks (BNN) on current benchmark tasks. The aim is to reduce the load of on GPU from arithmetic operations with huge matrices made of 32 or 64 floating point numbers.
Part-Time Lecturer
University of the Witwatersrand: 2019-2021
Lecturer 1st and 2nd year students the following topics: Introduction to Programming Using C++ and Python, Data Structures and Algorithms and Computer Networks.
Machine Learning Engineer
Retro Rabbit - Nedbank: 2019
Worked on Avo by Nedbank. I wrote the initial piece of code to do the customer segmentation logic to group the potential customers by spending habits. I was also very involved in the customer’s experience improvement by collecting the data from the customer from both the app anf the different sources to propose better products.
Machine Learning Engineer
Retro Rabbit - Kalido: 2018-2019
Worked on Kalido. My role was mainly to maintain and improve the different Machine Learning models for the systems on AWS Cloud. That was entailing reading new papers, A/B testing, custom dataset creation and training.