Our Work

Portfolio

Real solutions built for real problems — not demos or proof-of-concepts.

These are Techneek's own production systems — the exact infrastructure patterns we deploy for clients. We build for ourselves first, so every recommendation comes from something we've already been running for months.

Infrastructure

Production Infrastructure Build

At a glance
  • 99.9% uptime across all production systems
  • 100% automated deployments via Terraform & Ansible
  • Full environment rebuildable from code in under 1 hour

A fully self-hosted production infrastructure spanning multiple workloads — personal tools, an AI assistant platform, and a live business environment. All systems are interconnected through a secure private network, accessible remotely without exposing anything to the public internet.

The infrastructure runs dozens of services — from self-hosted productivity apps to a complete private AI stack — all protected by Zero Trust access controls. Automated daily backups keep everything recoverable with no manual intervention.

Challenge: Managing infrastructure across multiple physical locations with different hardware, different networks, and no centralized control plane. Manual deployments meant inconsistent configurations and no audit trail.

Solution: Standardized on Proxmox VE with Terraform for provisioning and Ansible for configuration management. Every server, container, and network rule is defined in code and version-controlled. Cloudflare Tunnels and Tailscale provide secure connectivity without exposing any ports.

Impact: Zero-downtime deployments, full infrastructure reproducibility from code, and a complete audit trail of every change. If a node fails, the entire environment can be rebuilt from scratch in under an hour.

Tech Stack

  • Proxmox VE
  • Terraform (bpg/proxmox provider)
  • Ansible
  • Tailscale
  • Cloudflare Tunnels & Access
  • Docker & LXC
99.9% Uptime Achieved
Daily Automated Backups
100% Automated Deployment
See IT Consulting & Infrastructure services →
Private AI & RAG

Private AI Assistant Platform — OpenClaw

At a glance
  • 10+ AI capabilities fully self-hosted
  • 0 bytes of data shared externally
  • $0 per-query cost after initial hardware setup

A fully self-hosted AI assistant platform built for complete data privacy. OpenClaw delivers private AI chat, speech-to-text, document analysis, intelligent search, and web-connected research — without sending a single byte of data to third-party APIs.

The platform uses LiteLLM as a unified gateway with intelligent routing — sensitive queries stay on-premise while general tasks use cloud APIs. Qdrant handles vector storage for RAG, enabling the AI to search and answer questions from private documents, contracts, and knowledge bases in real time. Document ingestion pipelines process PDFs, emails, web pages, and databases, automatically chunking and embedding content for semantic search.

Challenge: Needed AI capabilities (document search, summarization, chat, speech-to-text) without sending any data to external APIs. Commercial AI services require data to leave your network — unacceptable for sensitive business and client information.

Solution: Built a fully self-hosted AI platform running 10+ services: local LLMs via Ollama with LiteLLM routing, Qdrant vector database for RAG, Whisper for speech-to-text, and n8n for workflow automation. All running on a single server with GPU passthrough.

Impact: Full AI capabilities with zero data exposure. Every query, every document, every conversation stays on-premises. Per-query cost after setup: $0. The same architecture is now deployed for professional services clients handling sensitive data.

Tech Stack

  • LiteLLM (model routing)
  • Ollama (local LLM inference)
  • Qdrant (vector database / RAG)
  • Meilisearch (full-text search)
  • Mem0 (AI memory layer)
  • Speaches / Whisper (speech-to-text)
  • Apache Tika (document parsing)
  • Crawl4AI & Playwright (web ingestion)
  • n8n (workflow automation)
10+ AI Capabilities
100% Self-Hosted
0 Data Shared Externally
$0 Per-Query Cost
See Private AI & RAG services →
Monitoring

Business-Wide Monitoring & Alerting

At a glance
  • 24/7 active monitoring across all nodes and services
  • Alert response time under 60 seconds
  • Managed monitoring from $299/month

A unified monitoring and alerting platform that gives a single view across every server, service, and application in the environment. When something breaks or degrades, alerts fire automatically — before users notice anything is wrong.

Alerts are automatically routed to the right channel at the right time. A public status page at status.techneek.icu gives clients transparency into availability without having to ask.

Challenge: No visibility into system health across multiple servers and services. Issues were discovered by users, not by monitoring. No alerting, no dashboards, no log aggregation — just hope that everything was running.

Solution: Deployed a centralized observability stack: Prometheus scrapes metrics from every node, Grafana provides real-time dashboards, Loki aggregates logs, and Uptime Kuma monitors all public endpoints. Alerts route through n8n to the right channel within seconds.

Impact: Complete visibility across the entire infrastructure. Issues are detected and alerted on before users notice. Mean time to detection dropped from hours to under 60 seconds. The same stack is offered as a managed service starting at $299/month.

Tech Stack

  • Grafana
  • Prometheus
  • Loki
  • n8n
  • Uptime Kuma
  • Node Exporter
24/7 Active Monitoring
99.9% Uptime Achieved
< 1min Alert Response Time
See Infrastructure Monitoring & Management services →

Ready to Build Something Like This?

Every engagement starts with a free consultation. I'll review your current setup, identify gaps, and put together a clear proposal — no obligation.

Book a Free Call View Pricing