Wayfair
2021 — PresentSupplier Advertising
- Led architecture and rollout of an experimentation platform for display ads, defining per-surface holdouts and centralized decisioning across ad types.Enabled org-wide adoption of statistically rigorous experimentation with ~5% revenue tradeoff, accelerating iteration across multiple product surfaces.
- Drove evolution of auction ranking to incorporate long-term value signals (pCTR + pCVR), partnering with product and data science on strategy and implementation.Delivered +5.8% purchases and +6.7% ROAS, establishing a scalable framework for optimizing revenue vs. customer experience.
- Owned reliability and observability strategy for external ad integrations across the real-time bidding pipeline.Standardized latency/error attribution and monitoring, reducing incident triage time and improving detection of vendor vs. internal failures.
- Identified and drove resolution of a systemic ~60ms p99 latency regression impacting ad serving.Led cross-team and vendor escalation, isolating root cause and improving end-to-end latency visibility and accountability.
- Designed and implemented a high-throughput telemetry pipeline for ad auctions (request, response, winner events).Established a unified measurement layer for fill rate, latency, and experiment performance used by engineering, analytics, and product teams.
- Drove architectural improvements to the ad-serving gateway, redefining service boundaries, domain models, and removing legacy systems.Improved system reliability, reduced technical debt, and increased development velocity for multiple teams.
- Defined and scaled engineering standards across teams (testing, observability, architecture, code review).Increased consistency and code quality across shared infrastructure, reducing operational risk and onboarding friction.
Asset Management Platform
- Owned a distributed Kafka-based media ingestion platform supporting catalog assets and video processing across multiple systems.Improved observability and reliability of ingestion workflows, enabling scalable downstream consumption.
- Led migration to the next-generation ingestion system, coordinating rollout of 100% production traffic (~600K+ requests).Maintained ~0.1% error rate and stable latency while deprecating legacy infrastructure.
- Standardized API contracts and data models across ingestion services and consumers.Reduced integration inconsistencies and prevented downstream data quality issues across teams.



