Bhuvan D

Data analyst with 2 years at Google YouTube Monetization Trust & Safety. Built the SQL data model now used by the full-time analytics team as the single source of truth for escalation reporting to leadership, led an agentic analytics platform on Google ADK (pre-launch), and shipped a personal LLM project classifying YouTube policy violations across 14 categories at 75.8% accuracy.

Data Analyst, Trust & Safety at Google (YouTube Monetization) (2024-06 – Present)

• Developed and established the foundational data model for YTM Trust & Safety escalation analysis, which was subsequently adopted by the full-time analytics team as the single source of truth. This model provides a comprehensive view of the escalation lifecycle, forming the basis for all vertical leadership reporting.
• Architected and led development of a leadership-approved, reusable multi-agent analytics platform that scales insights generation across any Trust & Safety vertical by swapping a single context document. Designed the agentic workflow (Root → Database → Data Science agents), authored adaptable LLM instructions for SQL generation with a self-healing retry loop on query errors, insight extraction, and chart routing, and built an LLM-as-judge evaluation framework for output-quality benchmarking. Prototyped end-to-end on YouTube Monetization.
• Generate monthly trend analyses for YouTube Trust & Safety leadership tracking distribution shifts in violation categories across policy verticals; ongoing since January 2025. Reporting drives resourcing and policy-priority decisions for the next escalation cycle.
• Unblocked creator-revenue recovery for unlisted music videos and saved the Creator Experience team ~3 hours per reporting cycle - built an automated daily pipeline extracting data from an internal issue-tracking system where direct database access was restricted, feeding a reporting table used by the team for business reviews.
• Delivered the inaugural Merchant UX OpEx Dashboard managing 34–42% of 2025 UX budget and driving a 99% spend-to-forecast rate, replacing manual spreadsheet trackers for the UX Research team after successful A/B testing.