Research & Publications
I treat research as infrastructure. My published work spans AI-simulated UX research, the behaviour of LLMs on analytical tasks, and accessible NLP. The accessibility work translates English and Urdu media into Pakistan Sign Language for the deaf community.
Context-Free-Grammar (CFG)-based English/Urdu Translation System from Video & Audio to Pakistan Sign Language (PSL) Gestures
A multi-stage system that translates English and Urdu media into Pakistan Sign Language gestures. It uses a proposed context-free grammar for grammatical precision. The pipeline preprocesses audio, transcribes with speech/NLP, then parses into validated PSL sentences. I built it to lower communication barriers for Pakistan's deaf community.
Grounded Simulation: A First-Principles Architecture for LLM-Based Synthetic UX Research
The architecture behind Articos. It shows how to make LLM-simulated user research realistic, auditable, and defensible. You ground synthetic personas so findings can be traced to evidence and scored for confidence, rather than asserted.
Less Expertise, More Coverage: The Counterintuitive Effects of Prompting LLMs for Analytical Tasks
An empirical look at how the framing of expertise in prompts changes the breadth and quality of LLM output on analytical tasks. It digs into why more "expert" prompting does not always mean better coverage.
// SSRN · public link on release