Tech Blog
Technical notes from the workshop.
A running collection of Pacific Wharf posts on local AI systems, training runs, hardware limits, infrastructure, and workflow engineering.
Return to homepageNo posts match that search.
Infra / explainer
What Happens When You Run an AI Model Locally?
You type a prompt into a box and press Generate.
Open entry
Infra / models
Understanding Quantization Without Pretending I’m a Mathematician
If you download local AI models for long enough, you start seeing filenames that look like somebody mixed a model name with a firmware revision number.
Open entry
Training
First Local LoRA Training Run
Pacific Wharf's first completed local LoRA run, covering setup, runtime, checkpoints, and why proving the pipeline matters.
Open entry
Infra / limits
My First Real Encounter With VRAM Limits
Getting a local model running on the RTX 2060 felt like a breakthrough. Then VRAM introduced itself as the adult in the room.
Open entry
Local llm
Successfully Running a Local LLM on an RTX 2060
I just wanted to see if useful local AI workloads were actually practical on consumer hardware.
Open entry