Nov 13, 2026
LoRA & QLoRA: Inside the Update
When to fine-tune, how PEFT, LoRA and QLoRA differ, and where adapters fit inside attention and the MLP—with worked matrices, gradients, memory charts and a practical training plan.
Read post →Tagged With
fine-tuning
1 post connected to this tag.
Follow the work. Understand the machine.
I write about the maths, the kernels, and what happens when I try to make them work. Join me for the investigations, including the bits that go wrong.
ENGINEERING NOTES FROM ANTHONY SHIVAKUMAR
THE ILLUSTRATED FIELD GUIDE
How to think like an AI kernel & HPC engineer
- 10 pages to connect the pieces
- A real Whisper investigation
- A worksheet for your next experiment
KEEP IN TOUCH
A note when there’s something to share.
Get the illustrated field guide by email, followed by engineering notes as I write them.
Subscribe to emails from Anthony
First Name
Last Name
Email
Submit
- You’re joining the ShivasNotes email list. No fixed cadence.
- Your welcome email includes the PDF. Allow a few minutes, and check spam if it hasn’t arrived.
Need an intelligent system to work on real hardware?
Embedded systems · Robotics · Constrained AI · CPU and HPC · Accelerators · Distributed systems
Work With Us / Antshiv Robotics
ShivasNotes
Independent notes on software, systems, AI workflows, robotics, and building in public.
