i maintained the github repository for over 200 members, using javascript, php, html, and css to manage and update the codebase. i ensured that version control best practices were followed, keeping everyone on the same page and making it easy for members to contribute. my efforts not only kept the repository organized and up-to-date, but also continue to play a key role in enhancing the organization's digital presence.

i developed an innovative, open-source website designed to revolutionize the way researchers and practitioners interact with large language models (llms). leveraging powerful tools like shiny, plotly, and pandas, i created a dynamic platform that hosts and displays comprehensive data tables on over 1,000 llms. this project significantly enhances accessibility by providing an intuitive, user-friendly interface that simplifies the complex process of selecting the most appropriate llm for specific needs.

the website allows users to easily filter, search, and compare various models based on their specific criteria, such as architecture, performance metrics, and application domains. by integrating plotly, i enabled interactive visualizations that provide deeper insights into the data, allowing users to explore trends and patterns across different models. the use of pandas ensures efficient data handling, enabling real-time updates and seamless integration with various data sources.

this platform is not just a repository but a powerful decision-making tool that empowers users to make informed choices, ultimately contributing to the advancement of ai research and development. by making this resource open-source, i aimed to foster collaboration and innovation within the ai community, ensuring that valuable knowledge and tools are accessible to everyone, regardless of their background or resources.

i am working on revolutionizing app inventor by creating a new matrix learning block that combines gwt for a user-friendly interface with scheme (kawa) for backend logic. this innovation allows for efficient matrix processing and simplifies reinforcement learning algorithms, making these powerful tools more accessible for users.

additionally, i mentored 15 undergraduates, guiding them in integrating dall-e for artistic visualizations and chatgpt for real-time information retrieval. through hands-on support, i helped them enhance their technical skills and significantly improve their project outcomes.

i played (and am still playing!) a key role in creating the brand image for the startup, working closely with both the engineering and marketing teams to ensure that the product's identity aligned with our overall vision. collaborating with the engineering team, i helped translate the brand's essence into the app's design and functionality, ensuring a cohesive user experience. on the marketing side, i worked to develop strategies that effectively communicated our brand's message to our target audience, helping to build a strong and recognizable presence in the market. by bridging the gap between these teams, i ensured that our product not only met technical standards but also resonated with users on a deeper level.

connectify (click here to see its repository!)

mlh best networking hack

connectify is a robust web platform i developed with three teammates over a 48-hour hackathon, aimed at enhancing networking and collaboration within the tech community. the platform features a sleek and responsive frontend built with react js and tailwind css. we used firebase authentication to ensure secure sign-up and login processes, providing users with peace of mind when managing their accounts.

ruby the robot dog

affiliated with app inventor

i developed a reinforcement learning system to help ruby, my robot dog, get better at following commands like "sit" and "wave." i used python and the openai api to create a simple reinforcement learning system that tracks how well the robot performs these actions. to do this, i set up a matrix-based approach that helps me see how accurate the robot is and make improvements as needed.

for the user interface, i used kotlin to make it both intuitive and engaging. the interface features easy-to-use command buttons, so you can give the robot instructions without hassle. i also added speech-to-text functionality, letting you talk directly to the robot. to keep track of how well the robot is following commands, i designed a dynamic matrix that updates in real time. plus, there’s a snoopy image that changes based on the robot's actions, providing instant visual feedback and making the whole experience more interactive and fun.

block builder for rag

work in progress

i am currently, alongside an amazing team member, working on building an application that will aid users in making rag applications with drag and drop blocks, similar to scratch and app inventor.