Projects

VLI (Visual Language Model) anaylsis images using computer vision and an LLM to translate it to text

The way we used it in a robotic arm is we trained a robotic arm to pick up an object using machine learning

Once the arm learned to pick up one object we use a VLM to replace the object, in theory allowing for the arm to pick up any object we give it. By generating an image of it for the arm to see and describing what it is for the arm to pick up

Impliemented a 3 layer neural network from scratch and distributed the load across 4 GPUs using CUDA

No backtracking so its just the raw computation of the forward pass and the backpropagation.

Allowed for me to understand the underlying mechanics of neural networks. As well as how to allocate jobs to multiple GPUs.

Designed and built a circuit taking a dual channel audio input ranging from 0-16kHz.

Combinding the signal and sending them through 2 seperate low pass and high pass filters. Allowing for 2 different LEDs to light up based off the frequency components

Implemented a dual-tone multi-frequency (DTMF) decoder using Python and multiple IIR filters.

This project was a fun way to apply digital signal processing concepts and understand how telecommunication systems work.

Using Python and multiple IIR filters to decode dual-tone multi-frequency (DTMF) signals.

Developed a project for High schoolers had them place sensors in a custom wooden batbox to monitor the humidity, temperature, and light levels.

The data was contained on a microcontroller then using radio frequency communication the data was sent to a base station where it was stored and visualized on a dashboard.

Implemented a real time line following algorithm to a 4 legged quadruped dog using ROS2. Had a PID loop that would adjust the angle of the robot compared to the line and tried to keep it centered Using OpenCV for the image processing keeping the image grayscaled for simplicity. Overall fun project + points because I did it in Spain