Wahab Ahmad

During my final year at the University of Waterloo, I collaborated with five other engineers on our capstone design project. We developed "Slouch Sergeant," a device designed to help users improve their posture. This device attaches to your chest and monitors your posture in real-time. It records the data using an accelerometer and transmits it via Bluetooth to the Slouch Sergeant app on your phone. We engineered a custom printed circuit board (PCB) and created a 3D-printed casing for the device. The system provides feedback to help users develop better posture habits.

In addition to the posture-tracking device, we also developed a Flutter app to receive and display the posture data in real-time. The app not only shows your current posture but also records it for future analysis. To encourage good posture habits, the app features a rewards system where users can earn coins based on their posture performance. These coins can be used to purchase military-themed hats within the app, adding a gamified element to the experience.

Developed a web application with a Kotlin backend using the Spring framework and a React frontend for real-time play of past math contest questions. Leveraged socket connections to broadcast multiplayer events, while allowing users to filter questions by difficulty level for continued progress. Designed the application architecture and database schema, ensuring efficient data retrieval and minimal latency for a seamless user experience. Deployed the application on AWS using Terraform, enabling automatic scaling and cost optimization. Utilized AWS services such as EC2, RDS, and S3 to ensure high availability and data security.

I conceptualized and designed a systolic array module for highly efficient matrix multiplication operations. Through optimization and testing, I achieved a fast processing speed of 120 MHz for matrices of any suitable size. To validate the module's capabilities, I successfully deployed and tested the design on a PYNQ FPGA board (shown below), providing real-world proof of its functionality and impact

Built an autonomous car that incorporates line tracking, obstacle avoidance, and the ability to manipulate moving objects using a motorized mechanical arm. The car is equipped with a line sensor for tracking, an ultrasonic sensor for detecting obstacles and proximity, and a motorized arm for manipulating objects.

Have you ever wondered what generative models are? They are models that can create new data from scratch, whether it's audio, images, videos, or anything else you can think of. I had the opportunity to dive deep into generative models and create my own. Specifically, I designed and trained a generative model to learn facial features and generate new faces. The model was trained using grayscale images and was able to produce stunning, new, and unique faces. I'm proud of the work I've done and the creativity that went into this project. Take a look at some of the grayscale faces that were generated entirely from scratch:

Here are some colored faces that were generated from scratch:

I developed a LSTM neural network capable of predicting stock prices with accuracy. Using a combination of live tweets and stock prices, my model was able to leverage sentiment analysis and technical analysis to forecast market trends with precision. As an added bonus, I also created a user-friendly email notification system that alerted traders to optimal buying and selling opportunities. The green line at the top of the chart represents the Amazon stock price predictions made by my neural network, while the bottom chart displays Amazon stock sentiment data derived from the Twitter API.

I designed and trained a convolutional neural network for feature detection on facial images. As a test subject, I chose Kevin from the movie 'Home Alone' due to his distinct facial features. With the trained model, I was able to accurately identify and highlight Kevin's facial features, such as his eyes, nose, and mouth, with precision and speed.

I have experience in designing and building a RESTful API that revolutionizes task management. With the ability to perform CRUD operations on a MongoDB database, users can easily organize their tasks and stay on top of their workload. This API also features email reminders to ensure that users never miss an important deadline again. While the focus of this project was on the backend functionality, rest assured that the user interface was designed with simplicity and ease of use in mind.