Designed a system to process and ingest 80k+ medical documents, efficiently chunking and generating embeddings.
Developed a Retrieval-Augmented Generation (RAG) pipeline tailored for radiology reports, enabling clinicians to gain enhanced diagnostic insights.
Engineered an Orchestrator LLM to route queries across three specialized nodes (Retriever-only, RAG, and general-purpose LLM) for adaptive, clinician-friendly handling.
Established an evaluation framework using NDCG@5, achieving a score of 0.857 on 1K+ radiology reports.
Skills: Large Language Models, Retrieval-Augmented Generation (RAG), LangChain, Python, NumPy
Advisor: Dr. Ran Zhang - Assistant Professor, Department of Radiology
Autonomous System Research Intern
Organization: Nokia Bell Labs
Location: Murray Hill, NJ
Duration: June 2024 - Aug 2024 (2 Mos)
Created a multi-agent LLM system using LangChain and LangGraph for real-time problem-solving.
Explored and implemented Chain-of-Thought prompting to optimize reasoning and decision quality across agents.
Visualized and deployed the multi-agent system using streamlit for interactive demonstration and user testing.
Engineered test cases and executed experiments on different SOTA LLMs, ensuring high-quality optimization.
Skills: Large Language Models (LLMs), LangChain, LangGraph, Prompt Engineering, streamlit, NLP, Git
Manager: Dr. Thomas Woo, Research Group Leader - Autonomous Systems Research Department
Master's Research
Organization: University of Wisconsin-Madison
Location: Madison, WI
Duration: September 2023 - May 2024 (9 Mos)
Evaluated and ran comprehensive benchmark tests on various LLMs with different configurations on platforms with varying levels of hardware capability.
Assessed model safety and quality trade-offs, measuring accuracy, hallucination frequency, and toxic output in compressed LLMs.
Co-authored PalmBench, accepted at ICLR 2025, on compressed LLM evaluation for mobile use cases.
Skills: Large Language Models, Recurrent Neural Networks, PyTorch, NumPy, pandas, Git
Advisor: Dr. Suman Banerjee, David J. DeWitt Professor, Department of Computer Science
Associate Engineer - AI/ML
Organization: Qualcomm
Location: Hyderabad, India
Duration: July 2022 - July 2023 (1 Yr)
Introduced evaluation frameworks and metrics for ML models to enhance efficiency and accuracy by 10.26%.
Ran inference and benchmark tests using SnapDragon Neural Processing Engine, based on OEM requests.
Manager: Mr. Anand Chandrasekaran, Founder and CTO
Undergraduate Research & Teaching Assistant
Organization: Bright Academy (Previously Solarillion Foundation)
Location: Chennai, India
Duration: February 2020 - June 2022 (2 Yrs 5 Mos)
Lead the NLP group to translate the video input of a German sign language translator depicting the weather,
into cohesive and accurate German sentences.
We achieved 98.19% score retention in the ROUGE-L score and 86.65% in the BLEU-4 score,
while simultaneously achieving a 30.88% reduction in model parameters when compared to the state-of-the-art model.
Collaborated on an NLP project to classify unfair Terms of Service clauses using a
two-stage knowledge distillation approach with BERT.
Guided 5+ students in research, evaluated assignments and projects in Python and Machine Learning.
Oversaw the website development and managed the server for our research group.
Wrote bots for posting office hours and creating polls.
Advisor: Mr. Vineeth Vijayaraghavan, Director - Research and Outreach
Student Researcher
Organization: Sri Sivasubramaniya Nadar College Of Engineering
Location: Chennai, India
Duration: December 2020 - April 2022 (1 Yr 5 Mos)
Posited a novel architecture for Fake News Detection based on Transformer architecture, which considers the title and
content of a news article to determine its integrity.
Our work performed with an accuracy of 74.0% on a subset of the NELA-GT 2020 dataset. To our knowledge, FakeNews Transformer
is the first published work considering both title and content for evaluating a news article.
Proposed a robust and cost-effective automatic speech recognition model for the Tamil language leveraging Baidu's Deep Speech
architecture. Our work was compared against Google's speech-to-text API, outperforming it by 20%.