Onkar Apte

I build systems that automate real-world workflows.

Selected work

• Reduced system latency by ~30% by replacing polling with WebSockets.
• Built an execution engine that converts natural language into backend workflows.
• Designed data pipelines processing real-world video and sensor data.

Currently finishing my master’s in Computer Science at the University of Colorado Boulder.

I work on systems that need to be fast, reliable, and make decisions under imperfect data.

Curent focus

• ML-driven automation systems for operational workflows.
• Document understanding (pipelines) using retrieval-based methods.