projects

A selection of projects in NLP research (natural language understanding and reasoning), LLM safety, agentic AI, and LLM systems, ranging from production-ready platforms to academic benchmarks.

nlp research, reasoning & llm safety
project thumbnail

commonsense reasoning with logical entailment trees

developing benchmarks and evaluation methods for logical commonsense reasoning
project thumbnail

measuring model collapse under recursive summarization training

investigating the fundamental limits of recursive training and synthetic data proliferation in foundation model ecosystems
project thumbnail

lost in plot - contrastive learning for movie retrieval

dense retrieval system for tip-of-the-tongue movie search from vague descriptions

responsible ai
project thumbnail

algorithmic bias in llm-based depression detection

evaluating and mitigating fairness issues in language-based mental health classification
project thumbnail

medical ethics assessment of large language models

evaluating ethical reasoning capabilities of llms in clinical contexts
project thumbnail

resource-efficient llm fine-tuning for mental health support

parameter-efficient adaptation of llms for mental health conversation support

agentic ai & systems
project thumbnail

s-loraa: scalable multi-agent llm system

dynamic lora adapter orchestration for efficient multi-agent inference pipelines
project thumbnail

gradcompass - ai-powered graduate application assistant

multi-agent llm system for personalized graduate school guidance and application support
project thumbnail

scoutr: agentic football transfer intelligence platform

agentic ai transfer intelligence platform for club tactical and financial constraints
project thumbnail

text-to-comic: scalable system for transforming textual narratives into comic strips

end-to-end pipeline using GPT-3.5 and Stable Diffusion on GKE with Redis for high-throughput comic generation