Computer Science Graduate Earns IBM Fellowship for AI Research on Data Extraction
Naman Ahuja will graduate this May with a master's degree in computer science from Arizona State University's School of Computing and Augmented Intelligence. His research on converting unstructured text into accurate tables has earned him an IBM Infrastructure Master's Fellowship Award, which recognizes work with direct industry and real-world applications.
The problem Ahuja tackled is straightforward but persistent. Large language models can read and summarize documents easily, but they struggle when asked to extract precise information and organize it into structured formats like tables. Important details get missed. Information becomes inconsistent. Models generate claims unsupported by the original text.
In health care, this matters enormously. Clinicians conducting systematic reviews read volumes of research and extract key findings into tables for decision-making. The process is manual, time-consuming, and error-prone.
Breaking Complex Tasks Into Steps
Ahuja's solution, developed through his master's thesis, breaks the task into smaller stages. Instead of asking an AI system to generate a complete table in one pass, his approach first extracts atomic facts from the text, then builds a plan for table organization, then fills in entries incrementally.
This mirrors how a person would do the work: read carefully, decide what categories matter, populate the table piece by piece. The method creates what Ahuja calls traceability-the ability to verify each step and reduce errors.
His system is also designed for what researchers call living data, information that evolves over time. As new studies are published, the system updates existing tables instead of rebuilding them from scratch, maintaining consistency while incorporating new evidence.
Vivek Gupta, an assistant professor of computer science and engineering at ASU and head of the Complex Data Analysis and Reasoning Lab where Ahuja conducted his research, said the work reflects a broader shift in generative AI and LLM development.
"Naman's work really captures what we're trying to do," Gupta said. "We're focused on complex structured data, especially how to generate it and evaluate it correctly, so we can build AI systems people can trust in real-world settings."
From India to Industry
Ahuja completed his undergraduate degree in computer science in Hyderabad, India, before enrolling at ASU in 2024. At the university, he served as a teaching assistant for a graduate-level natural language processing course, delivered a guest lecture on neural networks, and presented his AI research at an international conference in Vienna.
He credits Gupta's mentorship with helping him navigate research setbacks. "Dr. Gupta has helped guide me through all the different tough aspects, which is important when experiments don't pan out the way you think," Ahuja said.
After graduation, Ahuja will join Amazon in Seattle as a full-time engineer working on large-scale systems. His focus remains on how models are built and how they solve real-world problems.
The need for this work will only grow. The world produces more text than ever. Someone still has to turn that text into usable knowledge. For now, that often means building tables by hand. Ahuja's work suggests they don't have to.
Your membership also unlocks:
