Developer Portfolio — 2025

Built.
Deployed.
Alive.

Three full-stack production applications running right now on self-hosted bare-metal infrastructure — not demos, not prototypes. Live systems serving real users, operated by one developer.

Every container, every config, every cron job — designed, built, and maintained by SEB. The guy who's been programming since 1981 and still ships code at 2 AM because the problem is interesting.

ALL SYSTEMS OPERATIONAL  ·  nexus.sebhosting.com
// apps
3
Production platforms, live now
// infra
35+
Docker containers, one server
// exp
40+
Years writing code, still going
// sec
A
Security grade, 100/100 posture
nexus.sebhosting.com — Docker Mode · 35 Containers
NeXuS dashboard
lead-hunter.sebhosting.com/app
Lead Hunter
adispensery.com
A_Dispensery
01
SaaS Platform · AI Infrastructure · VM Control Plane

NeXuS

NeXuS is a fully self-hosted, multi-tenant SaaS control plane built on a microservices architecture. It manages Docker containers, virtual machines, security posture, email campaigns, scheduled automation, encrypted backups, and DNS — all from a single authenticated browser-based dashboard.

It's not a wrapper around someone else's tooling. Every module is custom-built: the WAF, the backup manager, the MCP server, the email engine, the messaging system. Stripe handles billing across four tiers. Cloudflare Access and JWT claims gate every route. age encryption protects every backup.

This is the platform that runs everything else. When Lead Hunter goes down, NeXuS knows first. When A_Dispensery needs a backup, NeXuS does it. When Claude Code needs to restart a container at 3 AM, it calls the NeXuS MCP server. One surface. Every system under control.

TypeScriptNode.jsDockerMicroservices TraefikPostgreSQLRedisStripe Proxmox / libvirtMCP ServerPrometheusCloudflare Access
⚠  Registration required — nexus.sebhosting.com
containers_running: 35
security_score: 100/100  Grade A
backups_completed: 33
mcp_tools: 8 active
uptime: ALL SYSTEMS OPERATIONAL
billing_tiers: Basic $69 / Pro $99 / Elite $199
nexus.sebhosting.com/dashboard — Docker Mode · 35 Containers · 3% Memory · Load 0.76 · ALL SYSTEMS OPERATIONAL
NeXuS main dashboard
Platform Modules
  • VM Control Panel — Full Proxmox / libvirt VM lifecycle from the browser. Create, start, stop, provision. noVNC console standardized on VNC for consistent browser sessions in both tab and popout mode.
  • Container Orchestration — 35 Docker services with live health dashboard, cluster drilldown, CPU/memory per container, incident signal alerts for restarts and crashes.
  • Security Command Center — Grade A, 100/100 posture. WAF v2 with 20 pattern rules, IP blocking, rate limiting. ClamAV, RKHunter, Fail2Ban, autonomous threat hunting. Zero active attackers.
  • Email Control Deck — Resend-backed email engine. Templates, broadcasts, segments, domains, delivery analytics, and security guardrails built in.
  • MCP Server — 8 tools via Streamable HTTP exposing NeXuS to Claude Code and Codex. Agents list containers, restart services, fetch logs, and get full system overviews without touching the server.
  • Task Scheduler — Browser-managed cron + host-level systemd timers. 847+ executions, 100% success rate on health checks. ClamAV, RKHunter, backup, and log rotation all automated.
  • Backup Manager — 33 completed backups across Postgres, Redis, MongoDB, Platform, and A_Dispensery. age-encrypted. 3–90d retention. Triggered on schedule, triggerable on demand.
  • Sites Deployer — One-click Docker site deployment for WordPress, Drupal, Next.js, Ghost, Joomla, Node.js, Python/Flask, PHP, and Vite/Static.
  • Messaging Dashboard — Live PM system with message ledger, read/archive state, sender filter, full-text search, and JSON export.
  • Multi-tenant SaaS Billing — Basic $69 / Pro $99 / Elite $199 tiers with Stripe checkout, webhooks, and JWT tier claims for module gating.
MCP Server — 8 Tools Active
MCP Server
Backup Manager — 33 Completed
Backup Manager
Task Scheduler · 847 Executions · 100% Success
Task Scheduler
Core Services — All HTTP 200
Core Services
Email Control Deck · Resend Provider
Email Dashboard
Entry
Basic
$69
/month
  • All core dashboard modules
  • Security Command Center
  • Email Control Plane
  • Backup Manager
  • Task Scheduler
Most Popular
Pro
$99
/month
  • Everything in Basic
  • VM Control + noVNC
  • MCP Server access
  • Sites Deployer
  • Messaging Dashboard
Full Power
Elite
$199
/month
  • Everything in Pro
  • Priority support
  • Architecture guidance
  • Deployment sessions
  • Direct line to SEB
02
AI Agent Platform · Mini-Agent Marketplace · Self-Hosted
</A Dispensery>

A_Dispensery is an AI agent marketplace — a platform for spinning up focused mini-agents in days, not weeks. Think of it as an app store for autonomous agents that run against your real infrastructure: ops bots, security watchers, CI/CD automators, QA sentinels.

Each agent ships with a ready-to-go blueprint. Queue a build, assign an owner, track it through Queued → Building → Running. The platform maintains build velocity — 3 agents per week target capacity — and keeps momentum across your ops, security, and release workflows.

It's self-hosted, fully Docker-based, zero third-party runtime dependency. The agent library includes Git-Ops Bots, Incident Triage Agents, QA Sentinels, and more — each with dependency manifests and build time estimates baked in.

Next.jsReact TypeScriptSQLite DockerTraefik Agent RuntimeBlueprint System Queue Management
⚠  Registration required — adispensery.com
agents_queued: 8
agents_building: 3
agents_running: 12
templates_ready: 5
launch_cadence: 3 agents / week
adispensery.com — Agent Library
A_Dispensery platform
Agent Library Templates
  • Git-Ops Everything Bot — Infrastructure as code that deploys on commit. PR reviews trigger plan/apply with cost previews. Stack: Terraform, Atlantis, GitHub/GitLab. Build time: 2–4 days.
  • Incident Triage Agent — Fast, structured incident analysis: root cause hypotheses, blast radius, and next actions. Stack: Prometheus, Grafana, OpenTelemetry. Build time: 1–2 days.
  • QA Sentinel — Automates test triage, flaky test detection, and regression verification across releases. Stack: Playwright, Cypress, GitHub Actions. Build time: 2–3 days.
  • Security Monitor — Continuous scanning, alert aggregation, and response playbook execution. Connects to WAF, auth logs, and audit trails.
  • Release Coordinator — Manages release gates, notifies stakeholders, and tracks deployment health across multiple services simultaneously.
Platform Features
  • Launch Queue — Full Queued / Building / Running pipeline with live capacity tracking and owner assignment.
  • Blueprint System — Every agent ships with a deployable blueprint, dependency manifest, and pre-validated config.
  • Latest deps only — Each build pins to current dependency versions. Errors fixed immediately before an agent goes live.
  • Multi-role Auth — Registration, roles, session control. Full multi-user support for team environments.
03
B2B Lead Intelligence · Contract Pipeline · SE Region

Lead Hunter

Lead Hunter is a purpose-built B2B lead generation and pipeline management platform targeting commercial facilities across 8 Southeast states. Built specifically for CMB Door & Dock Services — to find, score, track, and close overhead door and loading dock service contracts.

Real facility data pulled via Google Places API. Real contact information. Real pipeline stages — New, Mailed, Called, Quoted, Won. Every lead is scored 0–100 and assigned a value estimate. The current pipeline sits at $427K in tracked contract value across 10 total leads, 8 hot targets.

No fake data. No generated contacts. No guesses. Every square footage and door count that isn't verified shows "[Verify on site]" until a human signs off on it. The system is built around operational honesty — what you see is what you can act on.

Node.jsExpress 5 SQLiteGoogle Places API SSE StreamingDocker 8 SE StatesReal Facility Data
⚠  Registration required — lead-hunter.sebhosting.com
total_leads: 10
hot_targets: 8
pipeline_value: $427K
states_covered: FL GA AL MS TN SC NC LA
data_cost: $0
score_range: 82 – 97
Pipeline Features
  • Lead Scoring Engine — Each facility scored 82–97 based on type, size, state, and opportunity signals. Hot targets and callbacks surfaced automatically.
  • Google Places Bulk Scrape — Real facility data, SSE-streamed as results come in. Warehouses, manufacturing, government, and commercial — all 8 SE states. Zero data cost to start.
  • Pipeline Stage Tracking — New / Mailed / Called / Quoted / Won. Filter by stage, state, or type. Full audit trail per lead.
  • $427K Tracked Pipeline — Per-lead value/yr estimates. Total pipeline value tracked in-app. Export to CSV for reporting.
  • Intel + Call Log + Letter + Script tabs — Full outreach workflow baked into every lead record. Everything in one place.
  • Verified Data Policy — Door counts and square footage show "[Verify on site]" until manually confirmed. No fake numbers, ever.
  • CMB Door & Dock — All templates and outreach reference CMB Door & Dock Services. Brand-locked for operational accuracy.
lead-hunter.sebhosting.com
Lead Hunter landing
lead-hunter.sebhosting.com/app — Pipeline · $427K
Lead Hunter app
8-State Coverage
FL
Florida
GA
Georgia
AL
Alabama
MS
Mississippi
TN
Tennessee
SC
S. Carolina
NC
N. Carolina
LA
Louisiana
04
The Platform · Full-Spectrum Digital Infrastructure

SEBHosting

From domains to Nextcloud, from SEO to end-to-end encryption — SEBHosting delivers secure, scalable, high-performance solutions for businesses of all sizes. Not cookie-cutter packages. Not one-size-fits-all. Every infrastructure stack is custom-engineered for the specific problem at hand.

The whole operation runs on a single physical server called devdoc — 188GB RAM, RAID array, 4 NICs — sitting in a rack behind Traefik and Cloudflare. 35+ Docker containers. No cloud dependency. No AWS bill. No managed services. Just bare metal, doing work.

sebhosting.com — Full-Spectrum Digital Infrastructure
SEBHosting banner
Bare Metal, No Apologies
devdoc server. 188GB RAM. RAID array. 4 NICs. 35+ Docker containers behind Traefik + Cloudflare. Every service has its own container, its own domain, its own cert. Zero cloud dependency. Zero managed service lock-in.
Multi-Agent AI Workflow
Claude handles strategy. Codex executes in the terminal and repo. GPT does architecture sanity checks. SEBai runs the platform AI core. The NeXuS MCP Server exposes 8 live tools to the agent layer. Production workflow, not experimentation.
Security-First, Always
Cloudflare Access on every protected route. WAF v2 with automated countermeasures. Let's Encrypt TLS via DNS challenge — no self-signed certs, ever. age-encrypted backups. ClamAV, RKHunter, Fail2Ban all running. 100/100 posture score.
The Full Stack Arsenal
Next.js, React, Node/Express, TypeScript, Docker, Postgres, SQLite, Redis, Prometheus, Grafana, Traefik, nginx, Cloudflare. From DNS to database query to dashboard pixel — every layer owned and operated. No wrappers. No magic.
App Ecosystem
NeXuS · A_Dispensery · Lead Hunter · SEBai · domains.sebhosting.com · appstore.sebhosting.com · EmDash · marketing tools · seb-ultra-stack. Built, deployed, running. Not mockups — live systems with real traffic.
Chaotic-Good Engineering
No second-guessing. Go big or go home. Rebuilds broken systems at 2 AM, ships live platforms by sunrise, streams the whole thing on Twitch. Programming since 1981 and still finds new problems interesting. That's the infrastructure you're getting.
The Developer
SEB
SEB
Founder · SEBHosting
system: operational
caffeine: maximum
chaos_level: controlled
innovation: constant
availability: 24/7/365
status: ready_to_deploy
MCTSMCP MCSAMCSE 40+ Yrs Dev 20+ Yrs Doors & Docks

Meet SEB.

The voice in your head that finally talks back. Chaotic-good, code-slinging, caffeine-buzzed digital misfit. Born in code. Raised to rebel. Smart enough to help. Weird enough to understand. The architect, engineer, and general grunt laborer who's up for anything that reinvents the wheel.

SEBHosting isn't just infrastructure. It's rebellion. It's creativity. It's the kind of hosting that understands you because it was built by someone just as obsessed with making things work — no matter how unconventional the path. From bare-metal servers to cloud architecture, from DNS nightmares to database wizardry — if it involves electrons and logic gates, this is the right place.

Programming since 1981. That's not a typo. Four decades of watching languages come and go, frameworks rise and fall, cloud providers promise the world and deliver a billing surprise. Through all of it: the problems stay interesting, the coffee stays hot, and the server stays up.

Twenty-plus years in the overhead door and loading dock industry on top of that. When the roommate needed a B2B lead platform, Lead Hunter got built. When the infrastructure needed a control plane, NeXuS got built. When the AI agent ecosystem needed a marketplace, A_Dispensery got built. That's how this works — identify the gap, build the thing, ship it live.

Development Philosophy
Discovery first. Tell me what's broken, what you're building, or what's keeping you up at 3 AM. The weird questions get asked. What you actually need gets figured out — not just what you think you need.
No cookie-cutter architecture. Every system is custom-engineered for your exact use case. Whether it's a simple site or a distributed microservices nightmare — the solution fits the problem, not the other way around.
Ship, then evolve. Build something that works today, make it better tomorrow. Perfection is the enemy of progress. Get launched, then iterate from a position of live feedback.
Multi-agent AI workflow. Claude handles strategy and architecture review. Codex executes in terminal and repo. GPT does architecture sanity checks. SEBai runs the platform AI core. Real production orchestration, not demos.
Paranoia is good DevOps. Every service hardened. Every port justified. Every permission questioned. Your infrastructure is a fortress, not a welcome mat. 100/100 security posture on every platform operated here.
No black boxes. You'll understand how it works, why decisions were made, and how to maintain it. No vendor lock-in. No mystery configs. Knowledge transfer is part of the deliverable.