Currently working as a Full Stack Developer at KariniAI
About Me
I began my journey in technology when I was in middle school, where I participated in VEX robotics. During quarantine, I read Neural Networks and Deep Learning by Michael Nielsen; I fell down the machine learning rabbit hole and started building projects with Python and JavaScript. I quickly realized my passion for coding and problem-solving, leading me to pursue a degree in Statistics and Machine Learning at Carnegie Mellon University.
Throughout my academic career, I've worked on various projects ranging from web applications to machine learning models. I'm constantly learning and exploring new technologies to expand my skill set.
When I'm not building, you can find me going on nature walks, reading murder mysteries, playing video games, or exploring new food places. I'm also passionate about music and I enjoy traveling.
Projects
MCP Registry™,an enterprise MCP portal integrated with KariniAI
- • Developed a centralized registry for MCP servers with security guardrails and governance controls
- • Integrated with KariniAI's GenAI platform for enterprise authorization and authentication, enabling usecases in agentic workflows in conjunction with native agent tools
- • Utilized Next.js, Tailwind CSS, and MongoDB for the backend
KariniANDI Chatbot,a powerful AI assistant for Microsoft Teams
- • Built real-time conversational AI bot with streaming responses, file processing with S3 integration, interactive feedback systems, and references display
- • Developed conversation state management, thread handling, and command-based interactions
- • Utilized Bot Framework SDK, Azure Bot Service, Node.js, and Karini.ai API
Ecosort,a Raspberry Pi powered Trash/Recycling sorter
- • Leveraged YOLOv5 trained on TrashNet to classify and sort trash & recycling via custom-built garbage cover powered by a Raspberry Pi 5, servo motors, and an ArduCam
- • Utilized multi-threading to streamline OpenCV frame captures with AI inference and used Firebase to display live updates
- • Able to separate water bottles, soft drinks, and other recyclables from everyday trash
PyTutor,a learning and debugging tool for students
- • Developed a learning tool that integrates a code editor, an interactive file explorer, and a recursive Python interpreter to enable line-by-line code-tracing
- • Supports complex functions, loops, and conditional statements
- • Implements OCR via PyTesseract to scan both typed and handwritten code from images
LieMap,a Netlogo model to simulate the spread of lies
- • Visualizes lie spread in different cultural contexts based lie severity, personal benefit, and social proximity
- • Utilizes NetLogo for visualization and parameterization of the model, which was based on literature review of lying dynamics
PokerAI,an advanced probabilistic poker bot
- • Developed a system that tracks opponent betting patterns and adapts strategy based on observed behaviors, including dynamically adjusted decisions based on range
- • Created betting strategies that balance risk-reward based on hand strength evaluation and opponent tendency analysis
Want to get in touch?
I'm always looking for new opportunities and collaborations. Feel free to reach out to me.