Improved benchmarking accuracy by building an ML-powered vehicle comparison tool using gradient boosting for price prediction and k-NN for competitor retrieval. Deployed the model via FastAPI and integrated into a PySide6 GUI to auto-generate PDF reports. Proposed and implemented an end-to-end workflow to streamline car and driver tracking in a testing unit.

Graduate Student NLP Researcher at University of Massachusetts Amherst

May 1, 2024 - October 8, 2025

Led the BioLongInstruct project to enhance Biomedical NLP by enabling long-context processing for extended medical texts via LLMs, including data curation for instruction tuning in Agile sprints. Implemented LongLoRA and other fine-tuning techniques on Llama2 and Gemma models to optimize performance on complex biomedical tasks. Built benchmarks and tuned LLMs across 10+ biomedical datasets using prompt and instruction tuning, achieving ~9% accuracy gains.

NLP Research Intern at L3Cube

August 1, 2023 - October 8, 2025

Conducted research on low-resource code-mixed languages, improving BERT-based language model performance by 10% through hyperparameter optimization with WandB. Coordinated dataset development, annotating 150k sentences for hate speech. Applied language augmentation and hybrid tagging techniques, yielding 10–20% improvement.

NLP Intern at iMocha

February 1, 2023 - October 8, 2025

Engineered an ML-driven Interview Assessment System using Speech-to-Text (Assembly AI) and trained 200-500 interview question samples for Azure LUIS-based scoring. Implemented a scoring system mapping 5–8 intents across 20 question types for precise assessment. Migrated project to the CLU framework in response to LUIS deprecation.

Python Developer Intern at Capgemini

September 1, 2024 - October 8, 2025

Improved benchmarking accuracy by developing an ML-powered vehicle comparison tool using gradient boosting for price prediction and k-NN for competitor retrieval. Deployed model via FastAPI and integrated them into a PySide6 GUI to auto-generate PDF reports. Brainstormed and introduced an end-to-end workflow to streamline car and driver tracking in a testing unit.

Graduate Student NLP Researcher at UMass Amherst

May 1, 2024 - October 8, 2025

Led the “BioLongInstruct” project, enhancing Biomedical NLP by developing long-context processing capabilities for extended medical texts using LLMs, including data curation for instruction tuning in an Agile environment. Implemented LongLoRA and other fine-tuning techniques on Llama2 and Gemma models, optimizing them for complex biomedical tasks and elevating their use for extensive healthcare datasets. Developed benchmark tasks and tuned LLM performance on over 10 diverse biomedical datasets using prompt and instruction tuning strategies, resulting in a 9% accuracy increase.

NLP Research Intern at L3Cube

August 1, 2023 - October 8, 2025

Conducted innovative research on low-resource code-mixed languages, strengthening BERT-based language model performance by 10% through hyperparameter optimization with WandB. Coordinated the development of a comprehensive dataset, accurately annotating 150k sentences for hate speech. Integrated language augmentation and hybrid tagging techniques achieving 10-20% improvement.

NLP Intern at iMocha

February 1, 2023 - October 8, 2025

Engineered an ML-driven Interview Assessment System using Speech to Text conversion via AssemblyAI and training 200-500 interview question samples from various books and online resources on Azure’s LUIS model. Implemented a scoring system interpreting 5-8 intents across 20 question types for precise assessment. Migrated whole project to newer CLU framework in response to planned deprecation of LUIS.

AI & Prompt Engineering Intern (Product Development & Implementation) at TRIP GOGO

November 1, 2025 - November 25, 2025

Improved generative accuracy of TripGoGo’s AI assistant using Retrieval Augmented Generation (RAG) with AWS Lambda, S3, and OpenSearch for scalable context retrieval and response optimization. Applied few-shot learning and prompt-tuning techniques to enhance response fluency, semantic consistency, and personalization across multilingual travel queries.

Machine Learning Intern at CAPGEMINI

September 1, 2024 - September 1, 2024

Improved benchmarking accuracy by 18% by developing an ML-powered vehicle comparison system using Gradient Boosting for price prediction and k-NN for competitor retrieval for client Stellantis. Deployed trained models via FastAPI, integrated them into a PySide6 GUI, and optimized the backend code to automate and improve PDF report generation processing time by 8 minutes. Designed end to end workflow to streamline car and driver tracking, enhancing test efficiency and reducing manual effort.

Graduate Student AI Researcher at UMASS

May 1, 2024 - May 1, 2024

Enhancing Biomedical NLP by developing long-context processing capabilities for extended medical texts using LLMs. Implemented LongLoRA and other fine-tuning techniques on Llama2 and Gemma models, optimizing them for complex biomedical tasks and elevating their use for extensive healthcare datasets. Developed benchmark tasks and tuned LLM performance on over 10 diverse biomedical datasets using prompt and instruction tuning strategies, resulting in a 9% accuracy increase.

AI Research Intern (Gen AI / NLP Focus) at L3CUBE

August 1, 2023 - August 1, 2023

Fine tuned and Instruction tuned Gemma (2B–22B-IT) models using LoRA PEFT, boosting F1 by 18% and evaluated generative quality via ROUGE, BLEU, AR-Hard/Easy metrics on English + Hindi data. Built and curated a 150K sentence labeled dataset for hate speech and profanity detection, enabling a comparative BERT study that improved model accuracy by up to 20% and enhanced robustness across multilingual domains. Published three peer-reviewed papers showcasing word and sentence level embedding evaluations and language augmentation techniques, achieving 10-15% gains in contextual representation and multimodal interpretability.