loading
reading
writing
saving
searching
WANTED Custom Linux AI Workstation Build Setup Local Project (NW Alexandria)
condition:
excellent
QR Code Link to This Post
This is a straight software and hardware staging project for a local builder who knows their way around a terminal CLI and deep learning stacks (Ubuntu LTS or Pop!_OS).
Project Compensation
- Flat-Rate Project Quote: Please submit a competitive flat-rate offer based on sourcing from your existing inventory (e.g., pairing a used tower base with a 3060). No hourly rates.
- (Note: This is a streamlined, builder-tier repurposing project focused on clearing idle stock; commercial enterprise pricing will not be considered).
Hardware Profile & The "Re-use" Strategy
This project is designed to clear out idle inventory. If you are a builder with Linux terminal CLI skills and older enthusiast rigs that lack Windows 11 TPM 2.0 support, this is your chance to monetize them. Drop in the specs below from your inventory, flash Linux, and configure the stack:
- GPU: NVIDIA RTX 3060 (12GB VRAM required for SDXL + ControlNet inference).
- CPU/RAM: Modern multi-core architecture (Ryzen 5/7 or Intel i5/i7) with 16GB+ RAM. (Older enthusiast rigs lacking Windows 11 TPM 2.0 support are ideal).
- Storage & Power: 500GB+ SSD, reliable 600W+ PSU from a reputable brand.
- Chassis: Clean, professional, or industrial Micro-ATX / compact mid-tower case. Zero aesthetic budget: No glass side panels, no RGB lighting, and no "gamer" styling. Needs to look like a standard office workstation on the outside.
- Connectivity: PCIe Wi-Fi 6 adapter.
Scope of Work & Deliverables
Source the hardware from your existing inventory, assemble, configure, and hand over a production-ready system:
- OS Provisioning: Clean, optimized installation of Ubuntu LTS or Pop!_OS.
- Environment Configuration: Proper installation of CUDA, Python, and all necessary dependency paths.
- AI Deployment: ComfyUI + Stable Diffusion backend setup.
- Model Integration: Pre-load SDXL base models and key ControlNet weights (pose/depth).
Live Evaluation & Purchase Process
Because this is a specialized headless AI node, a live demonstration is required using the assembled hardware before finalizing the purchase:
- The Setup & Test: Bring the system to my location (or arrange a local test). I will plug in my test suite, load 5 multi-source product images, and run a combined text-to-image and multi-image img2img upscale pipeline using ControlNet (1200×896 minimum resolution).
- The 60-Minute Burn-In: I need to observe the system successfully complete a continuous 60-minute stress test running heavy multi-image batches without crashing, thermal throttling, or VRAM errors.
- The Commitment: Once the system successfully passes this live validation on the spot, the project is officially signed off and paid out immediately.
How to Reply
If you have the hardware inventory and Linux expertise to tackle this, please reply with:
- Your background: Briefly note your experience with Linux environments and hardware staging.
- Preferred Linux distro: (Ubuntu LTS or Pop!_OS).
- Execution timeline: How quickly you can deploy this from your existing inventory.