Simon Bekezulu
Principal Data Scientist
Midrand
Summary
Meticulous Data Scientist with a proven track record of performance across variety of domains with unlocked revenue of R150m in FY23 leading a team of 5 direct reports at Vodacom.
Expert in the following areas:
- cloud computing
- machine learning
- deeplearning
- experimentations and design of AB test scenarios
- time series forecasting
- neural network
- LLM(Large Language Models, this implies prompt engineering and fine tuning open source models)
- Transformers (architecture underpinning LLM)
Overview
10
10
years of professional experience
Work History
Principal Data Scientist
Vodacom
03.2023 - Current
- Responsible for implementing data science strategy in Vodacom Business
- Lead all initiatives on implementation of data science capability from Vodacom Business and negotiate new initiatives with business
- Responsibility of up skilling the team of data scientists across the Vodacom Group
- Presented complex findings in a clear, concise manner using visualizations that informed executive decision-making processes.
- Optimized business processes for increased efficiency through detailed data analysis and visualization tools.
- Ensured compliance with industry regulations and data privacy standards, mitigating risks associated with sensitive information handling.
- Assessed accuracy and effectiveness of new and existing data sources, data analysis techniques and data models.
- Optimized machine learning pipelines and computational resources deployment strategies resulting in reduced processing times.
- Built company A/B testing frameworks to test model quality.
- Delivered revenue of R120m in FY23
Senior Data Scientist
Vodacom
11.2021 - 03.2023
- Built a product recommendation system for Vodacom business
- Adopted a malfunctioning churn model and automated and optimized for improved performance
- Adopted a deal matching framework and used the recommendation to tailor offering to the Vodacom Business customers
- Implemented Deal Scoring Engine in production
- Implemented NRV Bundle recommender model in production
- I have created the account level feature store and contributed features enroute to 10k feature target
- Implemented randomized sampling techniques for optimized surveys.
- I have mentored 3 data scientists, two senior data scientists and one junior data scientist
- I have assisted the business in sourcing and staging fixed products metadata features.
Senior Data Scientist
Ringier SA
09.2020 - 10.2021
- I am working on the group product propensity model
- This recommends products that customers are likely to buy both on the online platform and when they are nudged from the call centre
- POC has been conducted successfully
- I am also involved in a project where we are optimizing the channels through which marketing campaign are distributed.
Data Scientist
Discovery
05.2018 - 08.2020
- I have developed customer life time value model for the base of active rewards customer
- This was done in python
- I have lead the project on Big Data feasibility and lead a team conducting POC on different Big Data Vendors
- I have developed generic recommender system for Discovery that can be adapted by different business units and partners
- I have also developed a Vitality heart rate fraud detection algorithm, and conducted POC on whether to host on SDK, manufacturer’s firmware or Discovery App
- I have done AB testing of the recommendation system via email channel.
Data Scientist
First National Bank
06.2017 - 03.2018
- I have helped deploy the model and tune it in production which was for detecting fraudulent logins for website
- This was done in R
- I have been involved in project for detecting fraudulent transactions on customers accounts
- My role in this project was to work with Kalman Filter implementation to model customers transactions flow as signal that can be used to train the Kalman Filter on
- The objective of the project was to flag suspicious behavior and hand the results over to the forensic investigation team for further processing
- Tools, algorithms and objective of the project: LDA (Latent Dirichlet Algorithm for classifying customer transactions in cluster distributions) Kalman Filter for understanding normal customer behavior so we can flag anomalies
- Kafka message pipeline between jobs HDFS, postgresql Objective of the project is to be able to do all this in streaming with as low latency as possible.
Data Scientist
Standard Bank
03.2015 - 05.2017
- Text mining and sentiment analysis
Data Innovation Manager
Standard Bank
10.2014 - 02.2015
- I have been in the innovations department, where I was working on the strategic vision of the division as the innovation manager.
Software Developer
Standard Bank
02.2014 - 09.2014
- I have worked with a card tracking system called track and trace, I was implementing certain functionality and maintaining its stability, it was developed in java
- I have worked in Global Markets where I was a support analyst; I was supporting the equities desk and algorithmic trading, where we also had to optimize a system called quotes engine.
Business Intelligence Analysts
Massmart
2 2012 - 3 2013
- I have worked at Massmart in the graduate/internship program in 2012
- I was in the business intelligence department working as a business intelligence analyst
- My main responsibilities were to: Extract reports for the executive board of the Massmart and Walmart
- Extract data from the database using SQL
- Compile ad hoc reports
- Analyze trends in the business that might lead to some challenging issues
- Ensure high quality of the produced reports and make necessary amendments where necessary
- Liaise with relevant stakeholders to get the correct information about any issue that requires attention at any stage in the business.
Education
Higher Certificate - Foundational Management Certificate
Gibs
Johhanesburg 04.2001 -
Bachelor of Science - Applied Mathematics
University of The Witwatersrand
Johannesburg, South Africa 04.2001 -
Bachelor of Science - Computer Science
University of The Witwatersrand
Johannesburg, South Africa 04.2001 -
Bachelor of Science - Applied Mathematics Honours
University of The Witwatersrand
Johannesburg 04.2001 -
Skills
Neural Networks
Optimization Techniques
Excellent Communication
Intelligence Gathering
Interpersonal Communication
Goal Setting
Amazon Redshift
Agile Methodology
Data science research methods
NoSQL Databases
References
Mr MJ Lubisi, Director, VISA SA, 067 703 4709
Timeline
Principal Data Scientist
Vodacom
03.2023 - Current
Senior Data Scientist
Vodacom
11.2021 - 03.2023
Senior Data Scientist
Ringier SA
09.2020 - 10.2021
Data Scientist
Discovery
05.2018 - 08.2020
Data Scientist
First National Bank
06.2017 - 03.2018
Data Scientist
Standard Bank
03.2015 - 05.2017
Data Innovation Manager
Standard Bank
10.2014 - 02.2015
Software Developer
Standard Bank
02.2014 - 09.2014
Higher Certificate - Foundational Management Certificate
Gibs
04.2001 -
Bachelor of Science - Applied Mathematics
University of The Witwatersrand
04.2001 -
Bachelor of Science - Computer Science
University of The Witwatersrand
04.2001 -
Bachelor of Science - Applied Mathematics Honours
University of The Witwatersrand
04.2001 -
Business Intelligence Analysts
Massmart
2 2012 - 3 2013
Similar Profiles
LuckyBoy MashileLuckyBoy Mashile
Data Analytics Specialist at VodacomData Analytics Specialist at Vodacom
Innocent Mondli PhoswaInnocent Mondli Phoswa
Unbranded Channel Specialist at VodacomUnbranded Channel Specialist at Vodacom
MATIKO JOEL KISIRIMATIKO JOEL KISIRI
Manager-Fleet and Facilities at VodacomManager-Fleet and Facilities at Vodacom