📍 New York | ☁️ Cloud & DevOps Engineer | 🤖 AI Systems, Automation, and Utility Product Builder
AI Tools I work with daily:
I build automation-first systems across infrastructure, AI workflows, internal tooling, and standalone utility software.
My work keeps coming back to one idea:
find operational friction, then turn it into software
I'm building a multi-language MCP (Model Context Protocol) server ecosystem — Go, Python (FastMCP), and Rust implementations — with OpenAPI specs for each surface. Integrations target real enterprise platforms: Microsoft 365 (Planner, To Do, Loop), HPE OneView for bare-metal infrastructure management, Tanium for endpoint security, and Zerto for disaster recovery orchestration. The goal is agent-accessible infrastructure, not just chatbots.
A full production-grade automation platform I built from scratch:
- React 18 + TypeScript frontend with Monaco Editor (VS Code's engine) for in-browser script editing
- Django REST Framework backend with Celery + Redis for async task execution and cron scheduling
- PostgreSQL persistence with full audit logging and RBAC
- Enterprise SSO support: SAML 2.0, LDAP/AD, Azure AD OAuth, Okta, Google, GitHub
- Kubernetes-ready with Helm charts and GitHub Actions CI/CD
- Scripts execute in sandboxed subprocesses with configurable timeouts, CPU/memory metrics, and Sentry-integrated error tracking
This is not a toy. It's a platform for teams to manage, schedule, and monitor PowerShell and Python scripts through a hardened web interface.
A remote monitoring and management tool built across two layers:
- Go backend — fast, low-overhead API layer for endpoint communication and telemetry collection
- Flutter frontend — cross-platform dashboard for managing endpoints, running scripts, and viewing status
Built because commercial RMM tools are expensive and often overbuilt for the specific operational needs I work against.
Beyond the server implementations, the mcp repo has:
- Pre-commit hooks enforcing code quality across all language targets
- OpenAPI spec library for consistent interface documentation
- Structured contribution patterns for adding new platform integrations
A Kotlin + Gradle Android project with a Wear OS companion. Focused on a tight, specific UX — the kind of low-friction, haptic-first experience that standalone apps should have but rarely do.
A self-hosted lab running Kubernetes, local LLMs on GPU/NPU hardware, and infrastructure experiments that don't belong in a cloud account. It's where IaC patterns get validated, new tools get stress-tested, and nothing is sacred.
- K8SHomelab — live cluster: manifests, operators, Helm deployments, config iteration
- TerraformHomeLab — reusable IaC modules for homelab and cloud environments
- npu-windows — local AI model experimentation on Windows NPU hardware
| Layer | Tools |
|---|---|
| Cloud | Azure, AWS |
| IaC | Terraform, Ansible, Helm |
| Containers | Docker, Kubernetes (AKS) |
| Languages | PowerShell, Python, Go, Rust, TypeScript, Kotlin |
| Backends | Django REST, Go (net/http), FastMCP |
| Frontends | React, Flutter |
| Data | PostgreSQL, Redis, Celery Beat |
| AI/Agents | MCP, FastMCP, local LLMs, n8n, Antigravity, Codex, Claude, Perplexity, Gemini |
| CI/CD | GitHub Actions, Azure DevOps |
| OS | Windows (primary), Linux (servers + containers) |
| Monitoring | Sentry, structured JSON logging, health check endpoints |
Most of my active work lives in private repos. This includes:
- Internal automation systems and admin tooling
- Dashboard and portal prototypes
- AI workflow experiments and MCP orchestration concepts
- Product-stage wearable app development
- PowerShell-heavy systems administration tooling built for real operational environments
The public repos are a slice. The private side is where the systems actually run.
- Automation over ceremony — systems that eliminate work, not just document it
- Utility over novelty — tools that solve something real, not demos
- Productized infrastructure — if I'm running it manually more than twice, it becomes software
- Small tools that compound — narrow utilities that evolve into broader workflows
- Sandboxed, observable execution — everything gets logs, metrics, and timeouts
- Agent-orchestrated workflows using MCP as the interface layer
- Expanding enterprise platform coverage in the MCP server ecosystem
- AI for operational leverage: triage, summarization, anomaly flagging in infra pipelines
- Cross-platform utility apps (Wear OS / Android / Flutter) with tight, focused value propositions
- Local-first AI experimentation on GPU and NPU hardware
- 🌐 profile.quinnfavo.com
▶️ YouTube
Build the tool once. Remove the task forever.