navigate

hero_image
A Data scientist and Software engineer bridging the Gap Between Data and Technology.
Using AI and Code to create innovative solutions that solve real world problems

Building innovative aibusiness solutions

scroll down

“You can’t connect the dots looking forward, you can only connect them looking backwards. So you have to trust that the dots will somehow connect in your future.”
- Steve Jobs

It can be hard to trust in the process when you can’t see the bigger picture. But you never know what might be around the corner, so you have to keep moving forward. And one day, you may recognize that some of the hardest things you had to go through were also the best things that ever happened to you.


experience

Data Scientist – Pricing and Promotions

September 2022 – Present

  • I have collaborated with BCG Gamma consultancy over the past few years to build and implement state of the art models that have significantly improved the company's pricing decisions and revenue.
  • Worked with BCG to develop and refine a pricing model, including designing experiments to validate the model's accuracy and robustness.
  • Built and tested pricing models using statistical and machine learning techniques.
  • Conducted market research and analysed competitor data
  • Developed training materials and presentations to educate the stakeholders on the use and benefits of the models.

  • Data Scientist – Pricing and Promotions

    June 2021 – Present

  • Contributed to the company's success over the past few years by leveraging my technical skills and data-driven insights to solve business challenges for the largest retailer in Africa.
  • Analysed customer data to identify trends and patterns in behaviour.
  • Built and implemented predictive models to forecast demand for products and optimize inventory levels.
  • Monitored the performance and stability of data-driven systems and processes in production

  • Mobile developer – Flutter

    June 2020 - June 2021

  • Built and deployed 5 mobile apps and a react web app.
  • Designed and developed cross-platform mobile applications for iOS and Android
  • Worked closely with cross-functional teams, including designers, product managers, and QA engineers, to ensure that mobile applications meet business and user needs.
  • Implemented robust and scalable architectures for mobile applications

  • Data Scientist – Location Intelligence

    January 2021 - June 2021

  • Worked with GIS data and tools, including maps, satellite imagery, arcgis and demographics, to identify patterns and trends that can inform store location decisions.
  • Developed and implemented machine learning algorithms to analyse GIS data and identify optimal store locations based on factors such as foot traffic, competition, and demographics.
  • Created visualizations and dashboards to communicate the results of the store location analysis to stakeholders, including maps, charts, and tables.


  • “Experience is the name we give to our mistakes.”
    - Oscar Wilde


    my work

    01

    Savvy shopper

    savvy shopper mockup image

    Project Description
    Savvy Shopper is one of the largest projects that I have ever undertaken, requiring expertise in multiple programming languages, frameworks, and platforms. As the sole developer on this project, I was responsible for its design, architecture, and programming. The project took several months to complete and includes a web scraper component built using Scrapy and Scrapyd Server, cloud key-value store databases such as MongoDB and DynamoDB, a backend REST API server built with Flask, a frontend mobile app built with Flutter that can be deployed to iOS and Android, and a Chrome extension. This project also played a crucial role in my ability to secure my first job at Shoprite ZA as a data scientist, despite having a degree in Accounting.
    savvy shopper store feature
    savvy shopper image

    Mobile application features
    - The mobile application allowed uses to look at the prices of products from their favourite stores spanning over 3 different categories namely Clothing, Groceries and Electronic accessories and 12 different stores which where web scraped daily - For a chosen store the application would show you the products that were deemed to be "GOOD BUYS" or "BAD BUYS" based on an algorithm that looked at the historic prices of the products to determine if price of the products was cheaper/more expensive. - You could also search for a product as well if it did not appear in the list of "GOOD BUYS" or "BAD BUYS" and see its historic prices. - The application would also show you competitor's prices for the same product or similar products. - It also feature a "Shopping list" feature for groceries where you could create a shopping list and could search through different stores and compare prices.

    Tech Stack
  • Web crawler: Python, Scrapy framework, Scrapyd, Heroku(Hosting the crawlers), AWS Lambda(scheduling and triggering the crawlers)
  • Backend: Python, Flask,Flask restful, data science libraries like numpy and pandas
  • Database: Dynamodb, Mongodb
  • Mobile app: Flutter - includes ui, rest api connection to backend, state management
  • Chrome extension: Javascript, CSS, HTML
  • Chrome extension

    Description
    - After the success of the savvy shopper mobile app l went on further to develop a chrome extension based on the same data from the mobile application. - This was mostly meant for people who preferred to do their shopping on a computer.

    “Put yourself in a position to be lucky”

    02

    Google Lens Clone


    Project Description
    I was truly amazed the first time I used the Google Lens application. It seemed like magic - you take a picture and, regardless of how many other objects are in the frame, it quickly finds the object you're pointing your camera at and searches the entire internet for relevant information. Years later, I decided to take on the challenge of building my own Google Lens from scratch. It wasn't easy to figure out exactly how they had accomplished this, but I did a lot of research and studying, and eventually I had accumulated enough knowledge to build my own "Google Lens clone." This was one of my favourite "full stack machine learning" projects, in which I implemented multiple deep learning models within a full stack application and built the entire thing end-to-end with no help or assistance from anyone.

    Features
    - Point your camera at any object and snap a photo. - The application will take a picture send it to the backend and find similar images after extracting the most prominent object in the picture. - My version finds products which where scraped from the retailer GAME(https://www.game.co.za/) and returns the name of the product, price information, etc - the application allows you to also point using a bounding box which product in an image you want to search.
    Project architecture visualisation

    How it all works and the tech Stack
    - When a picture is taken and sent to the backend, the first ml model(Yolov7) finds the most prominent objects in the image and selects one with the highest probability. - An image clip of this object is created and is then passed on to the Resnet model which extracts features from the clip and turns them into embeddings of 2048 dimensions. - The embeddings are then queried against a precomputed index using Spotify's annoy library which returns the index of the best matches. - The index is then used to query a database of product information and the rest api returns this information. Tech Stack: - Devops: Docker, AWS EC2 Linux, GitHub version control, - Languages: Python, Dart(Flutter) - ML models: Pytorch, Yolov7, Resnet model

    03

    Mobile Ecommerce App


    Project Description
    - This is the first application that l have ever build for mobile. - l designed it from scratch taking inspiration from designs l found on dribble and then developed it using the Flutter framework. - l have found that l learn better by doing and its through applications like this that l am the developer and scientist that l am today. - This is my preferred way of learning

    Features
    - The application features a lot of different animations - It uses the provider library to manage state. - Whilst the application was mostly an MVP l has an authentication integration from Firebase as well as a payment gateway through Square.

    04

    Web Dashboards

    Black Friday Plotly Dashboard

    Black Friday Plotly Dashboard


    Project Description
    - Have you ever wondered if the prices you see on black friday marked as 50% off or 70% off are really discounts? - Well, you are not the only one, l had the same thought too and l decided to find that out thought this project. - I decided to put the largest retailer in Africa to the test and scrape their products prices. - The hypothesis was that some of the retailers would mark-up their prices just before black friday and then reduce them on blackfriday to make it seem like they have given customers a huge discounts

    Clothing Plotly Dashboard


    Project Description
    - This was build on the same stack and mind set as the blackfriday dashboard but l did a re-design and extended the scope to other stores. - Here is a list of things you can do on the app:
  • decide which products are over priced
  • which products are doing well
  • which ones are not performing very well
  • which ones is it trying to get rid of
  • which products to buy and when to buy
  • buy low, sell high

  • Tech Stack
  • CSS
  • Python
  • Flask
  • Pandas
  • Plotly/Dash
  • Clothing Plotly Dashboard.bmp

    React Dashboard


    Project Description
    - This dashboard shows my first attempt at the react.js library. - After l had become comfortable with flutter l wanted to move on to something in the web, given the versatility of website. - I chose the most popular frontend js library React.

    Tech Stack
  • React
  • 05

    Reflectly Clone


    Project Description
    - One of the applications l build whilst developing my skills as a mobile developer - Reflectly is one of the most popular journaling apps in the world. - This app helped me boost my confidence and reach out to some very cool developers. - This is one of the toughest mobile apps out their because of its complex ui and l felt that if l could develop something like this l would be able to do just about everything else.

    l am passionate about solving hard problems to drive positive change in the world.

    Building innovative solutions can be a challenging and rewarding process. It requires creativity, critical thinking, and a willingness to take risks and learn from failures.
    By following a structured process, such as the design thinking process, and gathering feedback from potential users or customers, l am able can increase the chances of developing a successful and impactful solution


    helperText


    Copyright 2023 | Project heavily inspired by https://caroselling.it/