MIKE XIAOSYS.VER: 0.9.4
<module id="whoami" />

What matters to me most in technology is whether it is genuinely useful in people's lives — whether it helps people live better, decide better, or build better.

My work spans both large-scale infrastructure and consumer product development, from distributed systems at AWS DynamoDB to full-stack, experiment-driven product work at Meta. Across both, I am drawn to problems where strong engineering, product judgment, and real human usefulness converge.

Experience Log

EXP_DATA

Software Engineer Current

Meta Platforms, Inc. / Facebook
Feb 2025 — Present

Led 0→1 product development for Facebook Friends Tab, a top-level surface serving 20M–25M+ daily active users in US/CA, spanning full-stack execution, experimentation, and AI-native workflow design.

Product & Experimentation
  • 0→1 Consumer Features: Drove multiple consumer-facing features from concept to production, owning architecture, full-stack implementation, experimentation strategy, and launch iteration across ambiguous product spaces.
  • Product Impact at Scale: Built and launched new product surfaces with experiments delivering +5.2% 0s DAU, +3.5% 15s DAU, +1.2% Friends Tab session count, and +7.8% high-quality comments sent to friends, contributing to +4.2% total Friends Tab interactions in US/CA.
  • Roadmap & Product Judgment: Shaped product direction through strategy docs, experiment readouts, and lookback analyses that influenced roadmap decisions and clarified not just how the team built, but what the team chose to build next.
AI-Native Execution
  • Prototype-First Workflow: Embedded AI-driven development into the team's workflow, reducing average prototype time for In-Feed Units from ~1.5–2 weeks to 2 days and helping establish a prototype-first model that made iteration faster and more intuitive for PMs and cross-functional partners.
  • Reusable Systems for Faster Iteration: Built reusable components, agentic workflows, and a Friends Tab-specific knowledge layer that compressed idea-to-prototype-to-experiment cycles from 2–3 weeks to under 1 week, increasing concurrent feature development from 2–3 features per month to 5–6.
  • AI-Native Team Tooling: Created internal systems including an Experiment Assistant interpreting 200+ metrics to accelerate launch decisions and a Code Review Assistant that performed domain-aware functional, quality, and risk review.
Technical Leadership & Decision-Making
  • Engineering Beyond Implementation: Expanded the role of engineering beyond delivery by improving how product, engineering, and leadership accessed context, evaluated tradeoffs, and aligned around faster, better-informed decisions.
  • Thought Leadership on Modern Development: Authored internal writing on AI-native engineering and modern development workflows that was amplified by leadership, shared across a broader working group, and adopted beyond the immediate team.

Software Dev Engineer

Amazon Web Services, Inc. / DynamoDB
Jul 2020 — Feb 2025

Built large-scale infrastructure for Amazon DynamoDB, driving reliability, scaling, and traffic management across a global storage fleet.

Distributed Systems & Reliability
  • GAC for DynamoDB Global Table MRSC (Tech Lead, Q4 2024): Architected and implemented Global Admission Control integration for Multi-Region Strong Consistency, enabling robust cross-region traffic management, accurate metering and billing, and zero-RPO support for DynamoDB Global Tables.
  • Split-For-Heat Protection: Designed and launched Split-For-Heat protection for DynamoDB's storage fleet, helping prevent unexpected throttling during sudden traffic surges and improving customer experience during periods of rapid scaling.
  • Fail-Open Traffic Admission: Designed and implemented a fail-open automation strategy for global admission control, preserving service continuity during CloudWatch outages and maintaining safe traffic admission when autoscaling signals were unavailable.
Technical Leadership at Scale
  • Fleet Health Service (Tech Lead, Q1–Q3 2023): Led development of a fleet health service monitoring infrastructure health across 250,000 storage nodes globally, and expanded the system across all 37 AWS commercial regions to improve visibility and operational response at worldwide scale.
  • DynamoDB Heat Insight (Co-Lead, Q3 2023): Co-led development of an internal traffic analysis service that generated heatmaps from customers' historical admission data, enabling AWS support engineers to diagnose throttling issues faster and reduce operational overhead.

// Archive_Logs

Software Engineer Intern
New York Life / Cloud Services
Jun 2019 — Aug 2019
Software Engineer, Gameplay
Forest Games / Beijing
Jun 2017 — Aug 2018

System Specs

Skills.sh
Languages
Java, Python, JavaScript, C#, PHP
Product & Full-Stack
React, experimentation, analytics instrumentation, 0→1 feature delivery
Systems
Distributed architecture, reliability engineering, REST APIs, traffic management
Cloud & Infra
AWS, DynamoDB, Lambda, API Gateway, CloudFormation, CI/CD
AI-Native
AI-assisted development, agentic workflows, rapid prototyping
Hardware Setup (Edu)

Southern Methodist Univ.

M.S. in Computer Science
2018 — 2020

Michigan State Univ.

B.S. in Computer Science
2013 — 2018

Data Streams

Input: Blog