I'm Liam, a dedicated developer passionate about creating innovative solutions and tackling complex challenges. With experience across diverse domains, I thrive on learning cutting-edge technologies and contributing to the developer community through collaboration and knowledge sharing.

I'm a curious and thoughtful individual who enjoys exploring new ideas, learning, and diving deep into topics that pique my interest. I value creativity, knowledge, and effective communication, often seeking clarity and precision in discussions. I have a penchant for understanding concepts and solutions thoroughly, which reflects my analytical mindset and dedication to getting things right.

Whether it's brainstorming, crafting detailed plans, or engaging in thoughtful conversations, I bring a mix of enthusiasm and a keen eye for detail. Always open to exploring fresh perspectives, I thrive in collaborative environments and enjoy the process of iterative improvement.

• 🌐 Experience:
  • Back-End Developer @ BoatsGroup: Summer and fall of 2024
  🔍 More about my work at BoatsGroup

  Focused on automating data validation for marketplace listings of boats and engines (makes and models). Designed and implemented robust systems to ensure data integrity using advanced validation techniques and automation tools.
• 📚 Education
  • Currently pursuing a Bachelor’s in Computer Science at RIT
• 🛠️ Exploring new technologies and pushing boundaries

The Hoyo Helper project simplifies the process of claiming daily login rewards for Hoyoverse games by automating logins to Genshin Impact (GI), Honkai: Star Rail (HSR), and Zenless Zone Zero (ZZZ). The standalone version supports bulk automation for multiple accounts, while the QT-based app provides a user-friendly GUI for storing accounts and configuring automated logins. Using Playwright, the project ensures efficient and asynchronous handling of login processes and session management. The backend scripting handles credentials securely, working seamlessly with the automation. Modular development allows each component—Playwright scripts, backend login functionality, and the GUI—to evolve independently while laying the groundwork for future integration. This approach ensures a cohesive user experience in the final product. Currently under active development, the project will eventually deliver expanded automation capabilities and an enhanced GUI. Contributions are welcome to help refine the code, report bugs, or suggest new features under the GNU Affero General Public License.

Check out the full project by clicking on the image here ^

This project demonstrates a complete Retrieval-Augmented Generation (RAG) pipeline, incorporating YouTube and web scraping tools to extract, process, and summarize content. The YouTube scraper downloads video transcripts and metadata, while the web scraper gathers text and links from specified domains, with both systems designed for scalable task management. Processed data is embedded using a custom API and stored in Milvus, a high-performance vector database, for efficient retrieval. The chatbot app utilizes Milvus to fetch relevant context and a generation API to construct conversational responses, presented in a Tkinter-based GUI with multi-user and multi-conversation support. The Milvus data loader ensures seamless embedding and storage of textual data with advanced indexing for optimized performance. Data visualization tools, accelerated by PyCUDA, enable efficient analysis of the retrieved content. Configuration is handled via environment variables and modular setups, allowing flexibility in scaling and customization. This project provides a foundational understanding of RAG pipelines and the associated technology stack for building sophisticated retrieval and generation systems.

Check out the full project by clicking on the image here ^