Vitalis User

Developed an agentic job-search assistant that parses resumes, scores candidate-job fit using LLM embeddings, and auto-tailors applications via a RAG pipeline over job listings. Implemented multi-agent orchestration with LangGraph-style research/writing/review roles and hardened the execution loop against tool failures (rate limits, 403s, malformed JSON) for graceful degradation.

Prototyped an agentic compliance and code review system that audits codebases and cloud configurations using static analysis plus LLM reasoning. Produced ranked findings with suggested PR-ready patches and generated structured JSON/Markdown outputs for SOC 2/ISO evidence collection and executive summaries.

Architected a high-availability k3s Kubernetes platform on AWS fully provisioned via Terraform/OpenTofu. Implemented GitOps delivery (ArgoCD), policy-as-code (OPA Gatekeeper), progressive rollouts, and automated node patching/drift remediation with Ansible AWX plus centralized observability.

Built and launched a voice cloning SaaS where users upload samples, fine-tune models, and generate studio-quality speech on demand. Implemented TTS/transcription using ElevenLabs/OpenAI and Whisper, built queue-based generation with Stripe billing and usage metering, and enforced strict tenant separation with Supabase RLS and signed-URL delivery.

Built a self-hosted LLM orchestration lab using n8n with a unified interface over OpenAI, Anthropic, and Google Gemini. Implemented RAG over Supabase and Pinecone using Tavily/Jina, connected React frontends hosted on Vercel to self-hosted n8n via secure tunnels, and hardened agent loops against external tool failures.

Defined and evolved an enterprise AI platform roadmap spanning experimentation, training, model registry, CI/CD, LLM serving, and observability across AWS, Azure, and GCP for healthcare workloads generating 8M+ monthly transactions. Delivered multi-tenant, multi-region high availability on EKS/AKS/GKE and improved GPU utilization by 47%, reduced AI response latency by 39%, raised contextual retrie

Launched a technical learning platform for DevOps, cloud, and AI engineering with curated labs, video content, and project-based tracks. Built a Jamstack frontend with serverless backend and used the platform as a testbed for Infrastructure as Code patterns, observability pipelines, and AI-assisted content generation.

Led MLOps and GenAI platform engineering for a regulated, multi-tenant AWS environment with zero customer downtime across releases, owning EKS GPU workloads, operators, storage classes, and Istio service mesh. Reduced annual cloud spend by $3.4M via GPU scheduling and autoscaling, cut infrastructure drift by 82%, and embedded compliance-as-code for HIPAA/PCI-DSS/SOC 2/SOX-regulated delivery.

Designed CI/CD and infrastructure automation for grid modernization across AWS and on-prem, migrating legacy deployments to Kubernetes with Helm-based releases and automated rollbacks. Improved availability to 99.99%, reduced MTTR by 67%, cut infrastructure provisioning timelines by 82%, reduced cloud operating costs by $3.1M annually, and accelerated CI/CD deployment velocity by 74% while strengt

Owned environment provisioning and infrastructure lifecycle management using Infrastructure as Code and database release automation, improving release readiness timelines by 65% and strengthening deployment governance for regulated ecosystems. Built enterprise monitoring/logging and observability using AppDynamics, Splunk, and ELK, and developed automation for WebSphere administration and patching

Completed a postgraduate degree program in cloud computing at the University of Texas at Austin.

Earned an HND in Architecture from Federal Polytechnic Nekede.