NLP & Generative AI Researcher
Making large language models more capable, reliable, and inclusive.
I research large language models and AI agents, from how we evaluate and align them to how we make generative AI work for underrepresented languages. I hold a PhD from UC San Diego, where I earned the Doctoral Dissertation Award and an Amazon Research Award, and have built language technology at Oracle, Grammarly, TikTok, and Meta.
31
Publications
17
Invited talks
43
Countries visited
Research & engineering at
Oracle
Grammarly
TikTok
Meta
Google Brain
Amazon Alexa
Adobe Research
Latest updates
“Agent-Testing Agent” accepted at EACL 2026 for Oral presentation
A meta-agent for automated testing and evaluation of conversational AI agents, with my intern Sameer Komoravolu.
arXiv ↗
Serving as EACL 2026 Chair for Student Volunteering
AraPIQA team wins Best Shared Task Contribution Award at the EMNLP 2025 MRL Workshop
Covering Arabic dialects across Morocco, Algeria, Tunisia, Egypt, Palestine, Syria, and the UAE, part of the Global PIQA shared task.
Global PIQA ↗
Started a new role as Principal Applied Scientist at Oracle
Released the Agent-Testing Agent preprint
Work done with my intern Sameer Komoravolu at Grammarly this summer.
arXiv ↗
Invited Talk: “LLMs: From Transformer Attention to Low-Resource Darija Summarization”
AI Forge Guest Speaker Session, World Learning Algeria 🇩🇿.
Invited Talk on Fast Prompt Alignment for Text-to-Image Generation
AI & Data Conf (Ramadan Edition), organized by Moroccan Data Scientists 🇲🇦.
Event ↗
Joined Grammarly as Applied Research Scientist
Team at TikTok released Fast Prompt Alignment for Text-to-Image Generation
Including the open-source codebase on TikTok’s GitHub.
arXiv ↗
Invited Panelist at the 2nd International e-Health Forum, Casablanca 🇲🇦
Under the High Patronage of His Majesty King Mohammed VI of Morocco.
Event ↗
Tutorial Speaker on LLMs at the 2024 Deep Learning Indaba, Dakar 🇸🇳
Also co-organized the Workshop on LLMs for Low-Resource Languages.
Event ↗
Industry Track Area Chair for EMNLP 2024
Invited Talk at the Workshop on Generative AI in the era of LLMs (I2COMSAPP, Mauritania 🇲🇷)
Talk ↗
Area Chair for ACL 2024
Team at TikTok released InfiMM-Eval
A benchmark for complex open-ended reasoning evaluation for multi-modal LLMs.
Project ↗
Joined TikTok (ByteDance Research) as Senior Research Scientist
Received the UCSD CSE Doctoral Dissertation Award for 2022 in person
Graduated with my PhD and joined Meta as AI Research Scientist
All opinions shared online are my own, and do not necessarily represent the views of my employer.