Work History

Building a decentralized datacenter powered by humans

• Pioneered a novel model sharding technique that allows LLMs to run across networks of smartphones
• Developed a differential privacy mechanism ensuring user data security while contributing to network computation
• Implemented a peer discovery protocol in Go that maintains efficiency on unstable mobile networks
• Scaled the beta to 200+ devices with performance matching dedicated GPU servers at 5% of operating costs

Measuring the effectiveness of State-Of-The-Art AI Alignment research

• Created a visualization framework for tracking alignment methodologies across major research papers
• Implemented a benchmarking system comparing alignment techniques against standardized safety metrics
• Developed tools to capture and quantify alignment regressions in evolving AI systems
• Built an interactive dashboard used by 30+ alignment researchers to coordinate testing methodologies