Spatial & Language Technologies Lab • Carnegie Mellon University
MSc in Engineering Artificial Intelligence, CMU '26
I am a Research Associate at the Spatial and Language Technologies Lab at Carnegie Mellon University. My research focuses on Human-AI Collaboration, Multi-Agent Systems, Social Intelligence in AI, and LLM Evaluation. As language models evolve from conversational chatbots into autonomous agents, effective collaboration requires far more than task completion—it demands understanding the unstated dynamics of interaction, inferring partner intent and beliefs, and navigating complex social environments. In this pursuit, I also investigate Computational Humor (generation, detection, and evaluation) as a vital element of natural human-AI interaction, developing cognitively grounded methods that make interactive agentic systems more robust, aligned, and trustworthy.
I build research frameworks and benchmarks for studying these problems in LLMs. I am the creator of HumorGen, a framework for cognitively grounded humor generation with 1,000+ model downloads on Hugging Face, HumorRank, and SIDE-Bench, a benchmark for evaluating social interaction and conversational behavior in multi-agent systems. More broadly, I develop reusable methods for LLM evaluation, social intelligence, reasoning, and understanding language model behavior.
Open-weight ecosystem for computational humor — six comedic personas, 14 LoRA adapters (EN · FR · ES), built with Cognitive Synergy distillation. Details in the HumorGen paper.
Tournament-based leaderboard web application for evaluating and benchmarking humor generation capabilities across modern LLMs. Details in the HumorRank paper.
End-to-end ML pipeline for multilabel detection of mental health conditions and cyberbullying using classical models and transformer-based embeddings.
Official portal for the CMU-Africa NLP Research Group — featuring a public research showcase, member dashboard, and collaboration system for African language technologies.
Specialized benchmark closing the "parametric knowledge gap" in AI economic analysis. 4,309 curated QA pairs from 220 World Bank African Reports (64,892 retrieval chunks).
Autonomous agent for business process automation using Google Forms/Sheets, LangChain, and FastAPI with integrated email notifications and QA monitoring dashboard.
Our paper Cross-Lingual Cognitive Synergy for Constrained Humor Generation in Large Language Models was presented at CLEF 2026 in Jena, Germany.
Our paper HumorGen was accepted to the Main Conference at EMNLP 2026.
Our submission won the CLEF 2026 JOKER Task 4: Humor Generation across English, French, and Spanish.
Started as an NLP Research Associate at the SaLT Lab at Carnegie Mellon University.
Gave a talk with CMU-Africa NLP on "Beyond the Transformer: How Modern NLP Research actually uses Foundation Models".
Mental health & cyberbullying detection paper received Best Paper Award at AAAI AIMedHealth 2026.
I am actively applying to PhD programs to advance research in socially intelligent AI that communicates with social awareness, reasons robustly across contexts, and collaborates effectively with humans—moving toward cognitively grounded machine intelligence.
