About Me

I’m Stephen Scarano, an MS in Computer Science (Data Science) and Bay State Fellow from UMass Amherst. I build learning systems that stay reliable when data is messy, biased, or asynchronous—from social-media polls to event-camera streams. My work connects measurement → perception → reasoning: debiasing online polling via post-stratification, fast optical-flow/egomotion for robust perception, and neuromorphic pipelines for low-latency object tracking. I’m currently a research engineer at the Noblis Autonomy Lab, supporting R&D for U.S. DOT and NASA. I’m applying to PhD programs (Fall 2026 entry) to pursue robust machine learning and neuromorphic approaches to trustworthy ML.

Selected papers:

• ICWSM (Twitter/X polls; bias & post-stratification) — paper [Finalist for Best Paper]
• ICCV’23 (optical flow / rotation in crowded scenes) — paper
• JQD:DM (Twitter/X polls; bias & post-stratification) — paper