Julian Ortega

Staff Engineer

• I wrote the product spec, designed the UI, and built a conversational AI platform for e-commerce support from scratch. I researched the problem by sitting with support agents, crawled 900 pages of knowledge base into a RAG pipeline, and iterated through three architectural patterns before settling on a single-agent-with-tools design that handled edge cases
• I built the evaluation infrastructure as its own workstream, validating 80% documentation coverage and a 30% ticket deflection target against real user questions before shipping
• I designed and built the architecture to migrate 500+ search servers from EC2 to Kubernetes, evaluating 3 architectural options and implementing the recommended event-driven approach with zero changes to existing data pipelines
• I reduced LLM inference costs by 20x through systematic model selection and experimental design across OpenAI, Anthropic, Gemini, and AWS Bedrock
• I drove CI/CD optimization delivering 3x build speedup at lower infrastructure cost across the engineering organization

Software Architect

• I architected a semantic search platform combining vector search with traditional search, adopted by enterprise retailers across 40+ languages. I designed both the backend embedding pipeline and the frontend search UI integration
• I led the technology transfer from an acquired AI company, serving as the primary technical liaison across onboarding, troubleshooting, and new feature development
• I founded the Software Architecture Group, establishing architectural governance, code quality analysis across 2M+ lines of code, and C4 documentation standards
• I built developer tooling and piloted AI code assistants with 15 engineers, improving developer experience across the organization
• I designed a 7-level dual-track career framework and curated a 45-book professional development index for the engineering organization

Team Leader / Line Manager

• I designed and built the Adaptive Personalization Engine from proof-of-concept to production across 4 major releases: a real-time microservices architecture with 6 core components processing 200+ user events per second
• I led the search modernization project for 2 years while wearing team lead, PM, and engineer hats simultaneously. I reduced reindex times from 2+ hours to 26 minutes and improved search response from 192ms to 130ms
• I built the search microservice frontend and backend end-to-end, achieving 84% test coverage and enabling enterprise-scale catalogues that opened new customer segments
• I mentored junior developers: 70% of team members received promotions during my tenure. I wrote role specifications and conducted technical interviews to scale the team

Software Engineer / Senior Software Engineer

• I built recommendation algorithms using collaborative filtering, processing terabytes of e-commerce data for international retailers. I created the evaluation framework during my internship that became production infrastructure
• I optimized the product catalog ETL pipeline for 32M+ product catalogues, reducing reindexing time from 40 hours to 4 hours through a data partitioning strategy I designed
• I led the migration from SVN to Git across multiple projects, improving team workflow and enabling modern development practices organization-wide
• I built a real-time recommendations engine proof-of-concept and integrated it into the core e-commerce product, establishing the foundation for the production recommendation systems that followed

Conversational Commerce Platform (Full-Stack, Self-Initiated)

• I self-initiated this project from research through PoC to production deployment, providing a narrative and live demo that supported new prospect engagement
• I evolved the architecture from modular intent routing to a single-agent-with-tools pattern, later validated as the approach adopted by leading AI providers
• I evaluated LLM providers and reached 20x lower cost than the initial vendor, with gains in intelligence, tool use, and speed

Search Suggestions System Redesign (Full-Stack)

• I reduced index processing times by 64-93% across customer deployments (e.g. from ~7.5 hours to ~30 minutes for large catalogues)
• I eliminated ~2 weeks of onboarding effort per customer by removing separate configuration for product suggestions
• I rolled the feature out to 10+ enterprise e-commerce customers through a structured beta program with side-by-side comparison tooling I built

Search Modernization (Full-Stack, Product Ownership)

• I reduced search reindex times from 2+ hours to 26 minutes across three release iterations
• I achieved 99.8% of queries under 200ms for mid-size catalogues, matching or exceeding the existing system
• I enabled enterprise catalogues with millions of items, opening customer segments previously beyond platform capacity
• I designed the hot-swap reindex strategy enabling zero-downtime index updates via Elasticsearch alias rotation

Master's in Computer Science (AI Specialization)

B.S. in Systems Engineering