Yixin Xiao
Senior Software Engineer at Capital One
5 years of experience building backend, platform, and cloud infrastructure systems on AWS. Strong background in Python, Terraform, Databricks, agentic AI tooling, and scalable internal platforms.
Experience
Senior Software Engineer · New York, NY
- Architected Terraform-based CI/CD pipelines to provision Databricks infrastructure and AWS resources for a platform serving 10,000+ users, reducing annual incident rates by 60%
- Designed and launched a Governance SDK adopted by 5 cross-functional teams within 2 months, standardizing Databricks platform integrations and reducing duplicate implementation effort
- Built an event-driven microservices system on AWS Lambda and EventBridge with automated health checks, reducing platform RTO from 10 hours to 15 minutes and cutting weekly operational support time by 93%
- Automated code auditing and sanitization workflows across the governance layer, eliminating 80+ hours of manual security remediation work per quarter
- Built an AI-assisted operational support agent and CLI workflow that tripled biweekly ticket throughput and cut new Lambda service setup from 1–2 hours to under 5 minutes
Software Engineer · New York, NY
- Optimized Databricks compute policies and workload guardrails, reducing platform spend by $100K per month ($1.2M annualized)
- Built a medallion architecture pipeline for Databricks chargeback reporting, enabling self-service analytics across 3 enterprise business verticals
- Led migration of 7,000+ users from legacy Private Cloud Databricks to SaaS Databricks, improving governance through centralized account management and stronger identity controls
Associate Software Engineer · New York, NY
- Built a centralized search portal for cybersecurity products and resources, consolidating 100+ entities used by 20+ internal teams
- Improved search relevance and query performance by up to 35% through TF-IDF wrapper optimization and AWS RDS schema improvements
Undergraduate Research Assistant · Ann Arbor, MI
- Built and deployed a full-stack search engine on Docker and OpenShift, integrating a Neo4j graph database with custom search algorithms to support partial and fuzzy matching of chemical compound structures
- Simulated chemical compound structures with Vis.js leading to interactive graph search methods, adding another search dimension
Education
BSE in Computer Science
Minor in Business Administration
GPA: 3.80 · Summa Cum Laude
Certifications
Projects
Multi-agent platform that orchestrates eight specialized AI agents — manager, designer, frontend/backend engineers, QA, deployment, monitoring, and user tester — through a 10-phase sprint workflow to autonomously build full-stack applications end-to-end. Supports web and mobile targets with configurable autonomy levels, user feedback integration via iterative sprints, and automated Playwright testing.
Containerized travel planning app with React 18, Express, JWT authentication, Docker, and Nginx with TLS. Automated development workflows using LLM-powered multi-agent orchestration with integrated code review and QA testing.
↑ built using the multi-agent framework above
Plant care management app with React 18, Vite, Node.js, Express, and PostgreSQL. Track plant collections with photos, set custom watering and fertilizing schedules, view a care dashboard with overdue and upcoming tasks, and get AI-powered care advice via Google Gemini. Includes email reminders, care analytics, and activity streaks.
↑ built using the multi-agent framework above
Serverless nutrition tracker on AWS (Lambda, API Gateway, RDS PostgreSQL, Cognito) with OAuth 2.0 PKCE authentication. End-to-end CI/CD with GitHub Actions and automated Playwright testing.
Web app for categorizing uncategorized liked songs into playlists with in-app playback and bulk assignment. Built with Next.js 15, React 19, TypeScript, Tailwind CSS, FastAPI, and Spotify OAuth 2.0.
⚠ this app uses spotify's api in development mode — users must be allowlisted to log in. see the repo README for details on how to get access.