Skip to content
View azterizm's full-sized avatar
⚛️
Reacting
⚛️
Reacting

Block or report azterizm

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
azterizm/README.md

Abdullah Memon

AI engineer. I build LLM systems that check their own answers, and I measure every change before I claim it. Python and TypeScript. Remote from Pakistan, UTC+5.

LinkedIn · X · memonsystems.com · abdullah@azterizm.com

Pinned Loading

  1. legal-rag-router legal-rag-router Public

    Sends a query to the exact legal source it cites before vector search, or refuses. No model, no network. UK legislation.

    Python

  2. legal-rag-verifier legal-rag-verifier Public

    Checks each sentence of an LLM answer against its source while the model writes it, and rewrites the ones that fail. Proof of concept, tested on UK law.

    Python 1

  3. blind-adapters blind-adapters Public

    Adversarial research. Reasoning adapters for an on-device legal model, fetched by PIR without a TEE. Adapter rule PASS, routing FAIL 10 of 12. Shows the limits a TEE removes. Builds on blind-counsel.

    Python

  4. blind-counsel blind-counsel Public

    Adversarial research. How far confidential legal AI goes without TEEs. Standard cloud RAG leaked 16 private tokens, the PIR design 0. Keyed model obfuscation leaked 100%. Shows the limits a TEE rem…

    Python