Open-Source Contributions & Personal Projects
- Contributed to llama.cpp (68k+ stars on GitHub), the leading open-source LLM engine
- Created callama, a Python interface for distributed llama.cpp deployment and management
Research & Development Projects
- Engineered customer service training simulator for PNC Bank using transaction-based persona generation
- Created a framework for systematic experimentation on LLM behavior patterns and cognitive biases
- Implemented word-of-mouth marketing simulation using multi-agent LLM
Creator of Pel; the First Programming Language for Language Models
- Designed a consistent, elegant syntax, optimal for transformer models training
- Implemented the Pel interpreter in Python as the host language
- Developed REPeL: Pel’s REPL (Read, Eval, Print, Loop), an interactive programming environment with advanced error handling features like
restarts
BEACON (Business Enhancement through Adaptive COordinated Networks)
- The first Agentic AI framework that enables agent orchestration and coordination all in Pel code
- Simulated firm departments through hierarchical specialized AI agents
- Provided wholistic, cutting-edge, and transparent insights to business owners
grAIdient
Developed an LLM-powered web-app to help students learn course concepts in an interactive way. This project was part of Prof. Anita Woolley’s course. We found out that using this app, students performed better in implementing course concepts in a real-world scenario, compared to using traditional case studies.
- 📰 Featured on the news! Read the full story here.
⛷️ asky
⛷️ asky is a macOS app that lets you interact with any text you see on the screen using the power of Large Language Models (LLMs). I’ve found it useful to stay in the flow
and here’s why:
Currently, every time you have a question about some text, you have to copy it, go to your browser, open a new tab, go to ChatGPT or Claude websites, paste the text, ask your question, and finally get the AI’s response.
⛷️ asky removes all the middle-men in the process: in any app you just select the text, press a shortcut, ask your question, and get the AI’s answer right there and then! No more distraction.