Experience
Industry
Multimodal Generative AI Research Intern GE Vernova
Developing multimodal generative models for wind turbine VR simulation, generating scene and asset content conditioned on physical and operational parameters. Building agentic workflows for power grid analysis pipelines, combining retrieval-augmented generation over technical documentation with MCP-based tool access to internal data systems.
Machine Learning Engineering Intern, Core Ads Ranking Reddit
Ran a multi-seed hyperparameter tuning study on Reddit's Progressive Layered Extraction purchase-conversion model, lifting click-through PRAUC by +3.2σ with better calibration. Extended the binary objective to multi-conversion counts via MaskedPoissonLoss, quantifying ranking vs. calibration trade-offs under train/serve caps.
Research
Graduate Research Assistant Center for Visual Computing, Stony Brook University
Conducting research in immersive analytics and human–computer interaction, focusing on large-scale XR environments and user-centered evaluation. Developing immersive visualization and interactive interfaces for the FlexiCAVE and SILO facilities, and LLM-assisted analytics frameworks and agents to analyze user behaviors in XR.