• Reaiverse is an achievement system that rewards every kind of real-life passion with an NFT. One can join as a player to earn NFTs or an NPC to issue NFTs to players

• Designed and built the first version of our user achievement ecosystem from scratch using Golang, MongoDB/MySQL, AWS SQS & Lambda

• Grew the tech from a single host service to an auto-scaling system on AWS using ELB (load balancing), ECS (auto-scaling), CodePipeline (CI/CD), and CloudWatch (logs, metrics & alerts)

• Grew the eng team from 0 to 6 and coached junior engineers on programming language, backend framework, database, and instrumentation via documentation and knowledge share sessions

• Being the home of the biggest Personal Identifiable Information (PII) at Uber (470TB of data), Driver docs team ensures reliable and secure user doc storage and processing. We build the right abstraction for multiple business verticals (Onboarding, Compliance, Driver/Vehicle Identity) to access Uber-verified user data

• Led 4 teams through a cross-functional 1-year project to fix the leaky funnel-switch between ride-share and food delivery. Driver activation went up by 48.9% and supply hours by 185% after roll-out

• Led a 6-month data remodeling project and produced a better abstraction for driver onboarding status. Coordinated with 10+ product teams to move to that V2 API

o Example impact: a product team managed to reduce the incorrect rejection of drivers by 8% after building on our V2 API

• Rearchitected team’s core service’s MySQL access pattern and increased read throughput by 7x (2.5k -> 18k rps), reduced P50 latency by 20x (500 -> 25 ms)

• Created and enforced unit & integration test standards for team’s core services

• Provided technical consultation to the senior leadership when defining the organization’s 3-year-plan

• Led a team of 6 on Python-to-Go rewrite of team’s core service and increased host efficiency by 6x (60 -> 10 hosts)

• Rebuilt an internal data backfill tool and increased write throughput by 30x (1 -> 30 rps), which cut operation agents’ manual work by ~4.5 hrs per week per person

• Standardized team’s inbound ticket handling procedure by creating a bot that managed ticket assignment, handoff and overdue reminder