Rockfish delivers a Synthetic Data Generation Platform for operationalizing AI workflows

Designed and shipped data ingestion and labeling pipelines that automatically process and categorize synthetic datasets, reducing onboarding time and automating end-to-end data workflows with dynamic model selection

Optimized data-generation pipelines with refined sampling and epoch scheduling, improving end -to-end data throughput by 10% + (based on data complexity) and rare-distribution coverage

Developed fidelity metrics and conducted competitive benchmarking against Gretel and Mostly AI, delivering a 12% increase in synthetic data quality

Curated Bluebonnet’s project repository - fixing broken links and replacing real voter data with synthetic equivalents - to deliver secure, ready-to-use examples for future cohorts

Developed SQL queries to extract demographic trends from the DNC BigQuery Warehouse

Embedded with political data campaigns for State Senate elections in Arizona’s District 9 and “Jeff Ettinger for Congress” in MN

Engineered voter distribution heat maps, elevating targeted outreach effectiveness by 30% versus the 2020 campaign

Blog: Data-Driven Doorknocking for Constituent-Centered Outreach Strategy (https://bit.ly/4fOpjzb)

● Developed Unity-based 3D VR assistive games to provide emotional support for 35 impaired teens

● Member of Website and Technology Design Team

● Led team that redesigned NSTEM's home page and key landing pages to improve visitor experience, content access, and comprehensibility

● Led a team that manages day-to-day website operations.

● Established new structured and interactive curriculum and weekly lesson plans for teaching online courses on Python and HTML at the San Jose Public Library