Software Engineer (CARSR)
B.S. in Cognitive Science (neuroscience/AI concentrations) and Computer Science
Software engineer with a dual background in Cognitive Science (AI/neuroscience concentrations) and Computer Science, and 2+ years of experience developing, debugging, and modernizing robust C++ and Python systems for high-stakes environments (FAA). Experience with automation, data-driven problem-solving, and rapid adaptation to complex technical systems. BlueDot Impact alumni (AGI Strategy) and CRS S-Risk Introductory Fellow. Deeply interested in understanding how complex/intelligent information systems work internally.
An experiment in recursive self-improvement: three AI agents collaborate on a task, then propose and vote on changes to their own coordination protocol. The scaffolding stays fixed; only the policy evolves. Starting from a minimal propose-and-vote loop, agents autonomously discover critique rounds, refinement passes, and eventually modify their own self-improvement process. Built with Gemini 2.5 Flash, structured output via Pydantic, and no frameworks.
Projects from my undergraduate years — kept as a record of early explorations in ML and systems programming.
A Python machine learning project using custom recurrent neural networks to predict future stock prices. Trained on 5200 days of IBM closing prices from the AlphaVantage API. Demonstrates time series prediction with detailed mathematical analysis of the underlying calculus.
A C++ project demonstrating third-party library integration with TCLAP for command line argument processing. The program reads text files, performs case conversion based on arguments, and writes results to output files. Showcases C++ file I/O, library integration, and shell scripting.