Gnaneswar Villuri Profile Picture

Gnaneswar Villuri Ph.D. Candidate

Stony Brook University

Contact: villurignanesh@gmail.com | Google Scholar | LinkedIn
Resume Web version + downloadable PDF.
View Resume Download Resume (PDF)

AI researcher with experience at Amazon AWS and Nokia Bell Labs, focused on building and deploying agentic, tool-augmented LLM systems that deliver measurable impact across production and research settings.

Research focus: Neuro-symbolic and tool-augmented large language models for reliable reasoning, with an emphasis on verification-aware prompting, reinforcement learning for tool-use policies, and agentic collaboration in open-ended coding and problem-solving tasks.

For a longer-form view of my research direction, see Research Overview.

Agentic LLM Systems Tool Use Verification Reinforcement Learning Neuro-Symbolic AI NLP

Publications

2025
Two Large Language Model-based Methods to Validate Open-Ended Problem Solving in Teams
Shaik, H., Villuri, G., & Doboli, A. (2025)
Artificial Intelligence Models and Systems Symposium (AIMS 2025)
An Overview of LLMs and a Novel, LLM-Based Cognitive Architecture for Solving Open-Ended Problems
Shaik, H., Villuri, G., & Doboli, A. (2025)
Machine Learning and Knowledge Extraction, 7(4), 134
PDF
Concept Combinations with Generator and Validator Agents Prompted Using Insights from Concept Networks
Shaik, H., Villuri, G., & Doboli, A. (2025)
International Conference on Complex Networks
Conference
An Experimental Study on the Interpretability of Transformer Models for Dialog Understanding
Villuri, G., & Doboli, A. (2025)
IEEE Conference on Artificial Intelligence (CAI 2025)
IEEE
A Stacked Multi-Layered Perceptron - LLM Model for Extracting the Relations in Textual Descriptions
Villuri, G., Shaik, H., & Doboli, A. (2025)
IEEE Symposium Series on Computational Intelligence (SSCI 2025)
IEEE
Towards Semantic Classification: An Experimental Study on Automated Understanding of the Meaning of Verbal Utterances
Villuri, G., Doboli, A., & Pallapu, H. R. (2025)
IEEE 15th Annual Computing and Communication Workshop and Conference (2025)
2024
An Overview and Discussion of the Suitability of Existing Speech Datasets to Train Machine Learning Models for Collective Problem Solving
Villuri, G., & Doboli, A. (2024)
arXiv preprint arXiv:2412.18489
arXiv
Using Speech Data to Automatically Characterize Team Effectiveness to Optimize Power Distribution in Internet-of-Things Applications
Villuri, G., & Doboli, A. (2024)
IEEE 3rd Conference on Information Technology and Data Science (CITDS 2024)
Automatically Understanding Human Behavior for IoT Applications with Optimized Human-in-the-Loop Control
Pallapu, H. R., Villuri, G., Doboli, A., & Doboli, S. (2024)
Towards Semantic Classification of Dialog using Contextual Prediction Networks
Villuri, G., & Doboli, A. (2024)
Cognitive Computational Neuroscience (CCN 2024)
Scholar

Experience

Graduate Research Assistant – Stony Brook University

01/2023 – 05/2027
Built LLM-based validation agents for collaborative coding, verification-aware generator–validator pipelines, neuro-symbolic dialogue/code representations, and tool-augmented LLM workflows with knowledge graphs and multimodal NLP.

Applied Scientist Intern – Amazon AWS (Mountain View, CA)

09/2025 – 12/2025
Evaluated the “GEPA” Automatic Prompt Optimization (APO) framework across 10 production LLMs (1B–405B), validating up to 12% structured reasoning gains and analyzing cross-model prompt transfer.

ML Research Intern – Nokia Bell Labs (Murray Hill, NJ)

06/2025 – 08/2025
Developed a multi-agent LLM pipeline using OPC UA for industrial telemetry insights; combined hierarchical agents, knowledge graphs, and function calling.

Software Developer Intern – Zippi Delivery (Stony Brook, NY)

05/2024 – 08/2024
Built multi-agent automation (“Zippi”) with CrewAI (95% routing accuracy), hybrid LLM setup cutting costs by 40%, and webhook automation handling 10K+ daily interactions.

Projects

SmartTutor: GraphRAG-based Learning Assistant

  • Developed an AI-powered tutor leveraging GraphRAG, replacing traditional RAG with a knowledge graph for improved accuracy of 34% and enhanced interpretability.
  • Processed 300 pages on Stereo Vision into a structured knowledge graph; achieved ~2s responses.
  • Enabled real-time cache visualization for transparent retrieval.

Language Neutralization System

  • Architected a real-time translation system for call centers supporting 10+ languages with 95% accuracy.
  • Constructed streaming pipeline (ASR → translation → TTS), reducing processing delay by 40%.
  • Orchestrated deployment via Docker and exposed a websocket API.

Agent Performance Platform

  • Led platform development to analyze 1,000+ daily agent interactions with text classification.
  • Fine-tuned model to attain 85% accuracy in identifying and scoring 15 distinct soft skill traits.
  • Executed MLOps practices, accelerating model update time by 70% and improving overall accuracy by 10%.

Open Source Contributions

GEPA — Configurable DefaultAdapter eval (score + feedback) · Dec 2025

  • Added a configurable per-example evaluator to DefaultAdapter and updated evaluation to return per-example score + feedback. PR #147

Skills

Python C++ JavaScript SQL PyTorch TensorFlow Hugging Face LangChain Docker SageMaker MLflow Neo4j Qdrant Pinecone

Contact

Email: villurignanesh@gmail.com  |  Google Scholar  |  LinkedIn