I'm Kieran, a certified Full Stack Web Developer specialized in creating robust E-Commerce applications. Here you'll find details on my projects, the technologies I use, and my personal interests. You can also check out my portfolio site at KC-7.com. Please feel free to check out my Portfolio Projects which I have linked below.

• 🌐 Focusing on personal project enhancement and development locally but I intend to share these with the wider GitHub Community soon.
• 🤝 Seeking collaboration with innovative developers and forward-thinking companies.

I leverage a diverse toolkit for both text and image generation, utilizing both local and cloud-based solutions.

Text Generation:

• User-friendly Interface: I employ TextGenUI alongside custom large language models (LLMs) sourced from Hugging Face, a popular repository for open-source machine learning models.
• Agent Teams: I also utilize AutoGen, a tool that allows teams of "agents" to be formed using either my local LLMs or ChatGPT, a large language model from OpenAI. To enhance the context these agents can access, I connect MemGPT to both AutoGen and the individual agents.

Image & Video Generation:

• User Interfaces: I primarily use Automatic1111 (A1111) and ComfyUI for image generation tasks. A1111 excels at image creation, while ComfyUI provides a user-friendly interface for specific models.
• Exploration and Integration: I have integrated Stable Diffusion with A1111 for AI video generation and explored other techniques like 3D modeling and image-to-text conversion. Notably, ComfyUI offers access to the Stable Diffusion XL Model with Refiner, which I've found to be particularly effective for local image generation, though it's incompatible with A1111.

Voice Manipulation:

• Pre-trained Models: I use So-Vits-SVC to manipulate voices using pre-trained models.

API Integration:

• OpenAI (Closed source despite the name): I have developed applications that leverage OpenAI's API for both text and image generation functionalities, showcased in projects like "AI Art Generation."
• Open Source LLM API: I have also experimented with making my own API by running an open source LLM locally (or cloud based using Google collab etc.) and then connecting to the API.

⭐ You can view my PyTorch Course Repo here!

• 5 ⭐ Project: E-Commerce Website - onlineAI.art

  • Languages, Frameworks, etc.: Django & Bootstrap
  • Code & README Link: Repo Link
  • Live Link: Live Link
  • Description: The E-Commerce Store project leverages Django and Bootstrap for an agile and scalable online shopping platform, offering AI-generated custom printed products. It features a user-friendly interface for seamless navigation and secure transactions, integrating Stripe for payments and AWS S3 for hosting. A unique pixel art avatar is created for each user, enhancing personalization. The site's admin section provides robust content management, and the codebase is maintained with diligent git practices for streamlined development and future updates.

• 4 ⭐ Project: AI Art Website - Cre8AI.art

  • Languages, Frameworks, etc.: Django & Bootstrap
  • Code & README Link: Repo Link
  • Live Link: Live Link
  • Description: This web application allows users to create custom AI generated artwork using the impressive DALLE-2 API by OpenAI. The generated image is automatically shared as a public post. The user can view it, change it to private, edit the description, download the image or delete the post. User accounts are limited to 5 generations per day. Users can also use the upload form to share a generation from a different site (example, stable diffusion). The site also has an About section which is managed from the site's admin panel, this allows admins to alter and create additional pages as required.

• 3 ⭐ Project: Computer Terminal Quiz

  • Languages, Frameworks, etc.: Python
  • Code & README Link: Repo Link
  • Live Link: Live Link
  • Description: This command-line application quiz app was built using Python to learn a wide range of different functions. The quiz takes the user's name, provides a series of questions and options and provides a final result at the end of the quiz. Each user input is validated to ensure the data is entered in the correct format. A google spreadsheet is linked to the app to keep track of the usernames and scores, the info is sent on completion of the quiz. The user will be given an option to print the high scores leaderboard to the terminal at the end of the game. The high scores are extracted from the google spreadsheet.

• 2 ⭐ Project: Car Web Quiz

  • Languages, Frameworks, etc.: HTML, CSS & JavaScipt
  • Code & README Link: Repo Link
  • Live Link: Live Link
  • Description: This is a fun, responsive car quiz that is designed to both engage and interact with the user. The site was designed to demonstrate a variety of different JavaScript functions. The user is asked a series of questions and their score is calculated in the top right-hand corner. The website interacts with the user and changes the colors of the body and the options to show if the answer was correct or not. A randomized message is displayed to the user after each answer alongside their result. The user is shown their total score at the end of the game and given the option to try again.

• 1 ⭐ Project: Life Hacks Website

  • Languages, Frameworks, etc.: HTML & CCS
  • Code & README Link: Repo Link
  • Live Link: Live Link
  • Description: The Life Hax website showcases responsive design, employing HTML and CSS for structure and style and is optimized across device sizes using media queries. Google Fonts and Font Awesome icons enhance aesthetics and user experience. Development followed best practices, with W3C validation and Lighthouse testing ensuring performance and accessibility.

• 👨‍💻 All of my projects are available at repos.kc-7.com

• 📫 Contact me at [email protected]