View System Configuration & Spec Sheet
talha@infer-workstation
-----------------------
Host: Custom Workstation (Fedora Linux / Kernel 6.x)
Role: AI Platform & Systems Integration Engineer
Focus: Local Model Serving, Build Toolchains & Linux Automation
Runtimes: Vulkan ICD, llama.cpp Compute Backends, GGUF Quantization
Langs: Python 3.12, Go, Bash / POSIX Shell, SQL
Infra: Docker, Podman, Tailscale, PostGIS, pgvector, Caddy
Status: Building deployable AI systems & automated infrastructure| Engine / Platform | Architecture & Scope | Core Stack | Status |
|---|---|---|---|
| VulkanIlm | Automated build toolchain and Python wrapper compiling llama.cpp against Vulkan backends. Enables local GPU inference on non-ROCm consumer hardware (Polaris/RX 580) and legacy non-AVX CPUs. |
Python CMake Vulkan GGUF |
|
| Goovie | Lightweight terminal-based streaming client in Go utilizing peer-to-peer data transport protocols with direct IPC piping to MPV. | Go WebTorrent MPV POSIX |
|
| Urban-Intel Engine | Containerized spatial intelligence engine orchestrating Vision Transformers on GeoTIFF satellite data with PostGIS topological verification. | FastAPI PostGIS Docker PyTorch |
|
| Buildables AI Track | Production LLM fellowship codebase: Hybrid vector retrieval, structured tool-calling, and automated CI evaluation suites. | FastAPI pgvector Docker |
|
Synced via local workstation session |
|




