Luis Norman

Software Development Engineer · Core Services Search

• Designed and built an event-driven data consistency service (SNS/SQS/Lambda) that decoupled search result validation from the request path, reducing p99 search latency by ~23% across all regions for 80M+ global listeners. The service processes 25K+ events/sec and triggers automatic reindexing when results diverge from authority sources.
• Owned a search diversification feature that fixed a position bias in click-based popularity by marking the most authoritative track per trackset at ingestion and switching ranking to label-provided signals. A/B tested over 14 days and improved search precision 7pp (52% to 59%) for 30M+ North American listeners with no recall regression.
• Re-architected Amazon Music's document ingestion system from a legacy self-hosted fleet to OpenSearch Ingestion (OSI), leading 2 engineers to build a router for index coordination and a pipeline-per-cluster design for multi-cluster fan-out and JOLT-based document transformation. Cut backfill indexing time by ~64% (18hrs → 6.5hrs) and introduced elastic auto-scaling.
• Migrated the Voice and Library platforms off a legacy monolithic core search service onto a new microservices architecture, coordinating across platform teams to update API contracts and resolve client compatibility issues.
• Built an internal admin tool for manually indexing documents to search clusters, used company-wide to resolve search visibility issues and reduce on-call burden.
• Serve as rotating on-call engineer for the core search platform, managing high-severity incidents for 80M+ global listeners.

Software Engineer

• Built an automated CI/CD fleet that provisioned clean APIC simulator environments with full fabric topology on every code review submission and delivered results via multi-channel notifications, shifting testing from a manual, once-per-sprint process to every PR with zero human involvement.

Software Engineer

• Joined as one of the founding engineers (4th hire), led design sessions with client stakeholders to define requirements, and built core mortgage lending automation software that helped scale the company from 4 to 20+ employees.
• Designed and built an SMS-driven mortgage automation service that allowed lenders to initiate and advance loan applications via text, creating a fully desk-free lending workflow with no manual processing steps.

Blockchain for Medical Records

Built a distributed system where processes compete across a network to mine and verify blocks (medical records) before being added to a ledger. Implemented from scratch using inter-process communication, cryptography, multithreading, and peer-to-peer networking.

IPC  ·  Cryptography  ·  Multithreading  ·  P2P Networking

Custom Memory Manager

Implemented a custom memory manager in C++ that replaces the standard library allocator, achieving a 3.5x speed increase through optimized allocation strategies and cache-aware data layout.

C++  ·  Memory Management  ·  Performance Optimization

Movie Recommendation System

Built a web application that recommends movies using collaborative filtering (user-based and item-based) and K-nearest neighbor, generating personalized recommendations based on a target user's ratings and their top-K similar users.

Collaborative Filtering  ·  KNN  ·  Flask  ·  Recommendation Systems

M.S. in Computer Science  ·  Full-time student, 2019–2021

B.S. in Computer Engineering  ·  Minor: Computer Science