I am a systems thinker and hands-on engineer focused on the intersection of Enterprise Architecture and Local AI. I specialize in modernizing mission-critical legacy systems and building scalable, automated pipelines.
A local, privacy-first pipeline for analyzing PBX call recordings.
- Tech Stack: Python, Whisper AI (Local GPU), Asterisk, Structured JSON Output.
- Goal: Demonstrating that high-value business intelligence doesn't require cloud SaaS.
Restoring and scaling a 12-microservice social network backend.
- Tech Stack: Java, Spring Boot, Docker, Kubernetes (k3s), CI/CD (GitHub Actions).
- Goal: Showcasing "Individual Amplification"—using modern DevOps and AI tools to manage a team-sized codebase solo.
My private R&D stack configuration.
- Key Features: Proxmox with GPU passthrough, n8n v2 workflows, Qdrant Vector DB, and Redis.
- Goal: Validating hybrid cloud/local LLM architectures.
- Languages: C# (.NET Expert), Java (Spring Boot), Python, SQL.
- Architecture: Microservices, Legacy Refactoring, API Gateway Design.
- DevOps: Kubernetes, Docker, Git/Bitbucket, CI/CD Automation.
- AI/ML: Local LLM Inference, Whisper STT, Prompt Engineering.
- LinkedIn: linkedin.com/in/leonid-zhuravel/
- Activity: I'm currently experimenting with Input Token Economics and Agentic Workflows.