Spearheaded cross-functional engineering initiatives spanning 5+ organizations to deliver the Meta Ray-Ban Display (TIME's Best Invention 2025). Defined the technical strategy and success criteria for 20+ experiences, establishing a culture of quality that aligned dispersed teams. Architected critical automation infrastructure, tooling, and metric guardrails to drive optimization across Performance, Reliability, and Efficiency pillars. Acted as the primary technical lead for high-stakes war rooms, resolving complex launch blockers to ensure the successful debut of a flagship hardware product.
- Led the development of AI-driven turn-by-turn navigation and local search by uniting 5 cross functional teams, filling data engineering and data science gaps to measure and visualize KPIs and driving surges to meet, saving successful executive demos. Wrote 0->1 bring up of maps and navigation product experience with product and design functions. Exceeded expectations on reliability, performance, and power via building tooling, analysis data, and writing high impact code. The experience is cited as "the best experience on device" by all major tech influencers and major publications
- Lead the performance/reliability/efficiency/quality program for the device transforming 18/20 failing product experiences across 5 organizations by architecting a quality assessment methodology, invested in tooling, and resolving systemtic issues on device impacting all teams
- Consolidated fragmented testing across 4 platform teams and 12 product teams, creating a canonical standard with 326 automation tests (functional, power, memory, performance), reducing test authoring time 85-97% and catching 14 SEV3 regressions during code review
- Authored novel concurrent-usage modeling that previous approaches missed, built tooling to measure for platform and 11 product teams to reduce memory-induced failures from 10%→0.2% user impact (goal was 2%), with this metric adopted as an org-wide top priority catching 2 critical production incidents
- Embedded in incident response rooms to root-cause architecture issues incorrectly blamed on network layer, redirecting 3 engineering teams to prevent weeks of wasted effort - improving voice assistant reliability 87.4%→92.7% and IG Reels video latency 63% reduction through data-driven diagnosis
- Five days before VP demo with blocking performance issues and no clear understanding, embedded in incident response to resolve all performance/memory issues (VP feedback: "snappy experience"), then architected frame rate guardrails that caught 19 regressions and 1 critical incident.
- Built network across platform, system, and product orgs to deliver/shape launch commitments, driving tiered KPI prioritization (P75 primary/P90 watchlist) and shaping org-wide KPIs
- Established formal mentorship program where a staff engineer now leads quality for a major product line, 2 mid-level and 1 junior engineer (promoted) shipped latency instrumentation and memory leak detection across 11 features, enabling org-wide capability scaling beyond individual contribution