I am in the process of developing a Neural Radiance Fields (NeRF) model from the ground up, inspired by the pivotal paper on NeRF available at arxiv.org/abs/2003.08934. This project is driven by my passion for cutting-edge technologies in robotics and 3D reconstruction, leveraging CUDA for accelerated computing to significantly enhance the model's performance. The objective is to create highly detailed and photorealistic 3D models from a set of 2D images, a capability critical for robotic vision and navigation systems.

The implementation involves:

• Designing a custom dataset loader for efficiently handling and preprocessing input images.
• Utilizing CUDA for parallel computation, enabling rapid processing of large datasets.
• Applying advanced machine learning techniques, including positional encoding and ray tracing algorithms, to capture the intricate details of 3D scenes.
• Integration with PyTorch for dynamic computational graph generation and backpropagation.

A cutting-edge music generation and visualization platform, created by Maximelio Finch and Sirena Backham, leveraging AI and web technologies to craft a novel musical experience. The project features a server-side API for music generation and a client-side application for dynamic 3D visualization.

• Developed a Node.js backend with Express for robust server-side scripting and API processing.
• Integrated OpenAI & Replicate for AI-driven music creation, enhanced by FFmpeg for media processing.
• Built a reactive client-side UI using React and animated 3D graphics with Three.js for music visualization.
• Utilized Vite for efficient project building and included tools like eslint for code quality assurance.
• Incorporated AI models for content generation, enabling users to create music and album covers from prompts.

Developed an AI-assistive drawing application designed to enhance artists' workflows, facilitate art learning, and accelerate art production

• Integrated a Style Transfer TensorFlow model and Stable Diffusion (SD) models as widgets, with processing managed by CLIP Drop’s SD API>
• Crafted and refined the drawing canvas, incorporating industry-standard controls for enhanced user interaction
• Engineered the program for multithreading, ensuring each AI widget operates on a separate thread for optimal performance.
• Actively working on the development of a DALLE2 plugin for advanced spot correction

• Trained a diffusion model on Google Colab to produce images and edit input images of myself in the art style of the BLEACH: Thousand Year Blood War anime.
• Converted model weights to a checkpoint file (ckpt) to be used in web-UI's. • Created full AI Animation with frames spliced using OpenCV/Python.

Utilized OpenCV's pre-trained Cascade Classifier to detect, frame and label cars in motion on a freeway.