At Wizybot, I lead the AI platform’s technical direction, from architecture and infrastructure to engineering standards. I reduced inference costs 80% through cross-provider benchmarking and prompt optimization, and cut database infrastructure costs 50% through query optimization and instance rightsizing.
As a Founding Engineer at Wizybot, I built the core AI response engine and retrieval pipeline for e-commerce conversations. I also developed product search and ranking, multimodal image and video recognition, and integrations for Shopify, VTEX, and WooCommerce; migrating core sync to queue-based processing improved throughput 20x.
At Neural Data Science, Columbia University, I’m modeling how neurons across the mouse brain encode decisions, sensory inputs, and movement using the International Brain Laboratory’s Neuropixels dataset. My background also includes theoretical biology research and hands-on machine learning work, from implementing algorithms and neural networks from scratch to building production systems.
