At Uber, I build Java-based microservices for Human-in-the-Loop and automated data labeling platforms. I integrated Automated Speech Recognition providers, cutting transcription handling time by about 37% and saving $15,800 per month.
I led the Smart Sampling rollout, reducing review costs by about 68%, and worked on uLabel Lite, a cloud-native version of Uber’s labeling platform on Google Cloud Platform. I also developed a model deployment framework using Vertex AI and Cloud Run for client-specific data requirements.
Earlier, as a Software Engineer Intern at Uber, I architected APIs for externalizing uLabel and streamlined client onboarding. At Aomics GmbH, I engineered a real-time trade forecasting pipeline and evaluated time-series models, reducing MSP Error to about 0.54% with N-BEATS.
