Luke
MacNeil

Artificial intelligence systems

• Work at Sandia is not publicly describable. Details on request where clearance and classification allow.

Cloud infrastructure & AI systems · 10 properties · Remote

• Architected and operate all production AWS infrastructure (EC2, Lightsail, SES) supporting 10 properties. Defined SLOs with error budgets for booking platform availability — 99.99% measured uptime — plus synthetic monitoring, runbooks, and incident response procedures.
• Saved $50K+/year by designing and shipping a custom multi-department CRM (Django, Celery, PostgreSQL, Redis) replacing paid SaaS: sales pipeline with Kanban boards, investor relations, multi-queue ticketing, hiring pipeline, and AI content publishing.
• Engineered dual-database architecture bridging the CRM (PostgreSQL) with legacy hospitality MySQL over SSH tunnels with health checks and graceful lifecycle management, enabling real-time guest enrichment across properties.
• Defined the company technology roadmap with department heads; led build-vs-buy evaluations, choosing custom builds where ROI justified them and SaaS where it did not.

Infrastructure, AI systems & full-stack development · Remote

• Designed a resilient data ingestion pipeline (CareerPulse) across 14 external APIs with independent failure modes — per-provider circuit breakers, adaptive rate limiting, SHA-256 deduplication. 1,024 automated tests with parallel GitHub Actions CI.
• Built production real-time infrastructure (Luke at the Roost) handling 10 concurrent WebSocket audio streams with 5-provider LLM failover, concurrent async recording across 7 audio channels, and automated multi-platform publishing. 58 episodes in production.
• Engineered multi-process IPC architecture (Hivemind) with isolated tmux sessions, zero-copy output streaming with automatic degradation, and context-isolated security boundaries. Multi-platform builds across macOS, Windows and Linux. 632 tests.
• Designed reusable GitHub Actions CI/CD templates with parallel test matrices across all projects; implemented a git worktree workflow enabling parallel agent-driven development streams without branch conflicts.

Microservices migration · Puppet · Government Cloud (FedRAMP)

• Spearheaded migration of monolithic infrastructure to microservices on AWS and GCP using Terraform, Packer and Kubernetes, cutting environment provisioning from days to under an hour.
• Engineered and maintained Puppet configuration management for 10,000+ production servers including Salesforce Government Cloud, maintaining FedRAMP compliance. Authored configuration standards adopted across the infrastructure org.
• Led a 5-person engineering team as Scrum Master and technical lead, reducing manual deployment time by 40% through CI/CD automation.
• Architected monitoring across 10,000+ nodes (Nagios, Zabbix, Zenoss) with custom alerting; established on-call rotation and postmortem process that reduced repeat incidents.
• Invented a machine-learning photo auto-tagging system using Salesforce Einstein AI — successfully patented.

Enterprise Linux · nationwide retail infrastructure · PCI-DSS

• Orchestrated nationwide migration of 7,500 retail servers from SCO UNIX to SuSE Linux Enterprise Server on IBM hardware — on time, with zero business continuity impact.
• Designed core infrastructure services (DNS, SVN, NFS, MySQL, Apache, Zabbix, CUPS) for a fleet of 15,000+ hosts with 50,000+ monitored items.
• Led PCI-DSS compliance for retail pharmacy infrastructure, passing all external audits.
• Developed a centralized automation platform managing software distribution, drive encryption and printer configuration across 7,500 retail endpoints.