PhD Student · Agentic AI for Scientific Discovery
My research focuses on agentic AI for scientific applications. I believe AI should possess reasoning and decision-making capabilities grounded in scientific principles to accelerate discovery across disciplines.
A modular framework for constructing multi-agent systems for retrosynthesis planning under multiple objectives.
Agentic system integrating evaluation directly into retrosynthesis planning for constrained tasks.
Agentic system for automating key stages of preclinical drug discovery through language-guided reasoning.
Offline-online reinforcement learning framework for improving synthon completion in retrosynthesis workflows.
Web-based platform for interactive visualizations of amino acid coevolution.
Deep learning models for predicting metal-binding sites in proteins from amino acid coevolution data.
Unsupervised analysis of gene expression changes in drugs to identify etiological patterns for drug-induced lupus.
Simulation testbed configurable via GUI for evaluating coordination across multiple autonomous platforms in complex environments.
Data analysis pipeline for anomaly detection and distribution shifts in health signals from heterogeneous sensor inputs.
Protocol for finetuning and evaluating language models to align with human annotations, accelerating research workflows.
Bioinformatics platform for visualizing and analyzing metagenomic data.
Instruction-tuned large language model for chemistry tasks using large-scale curated datasets.
Benchmark for evaluating language agents in scientific discovery tasks.
Email: lastname dot 3239 at buckeyemail dot osu dot edu