Gaurav Keshari

About

Hello! I am B.Tech student in Artificial Intelligence & Data Science from NMIMS University graduated in August 2025, my professional experience includes a role in Research Lead & AI Develpoer at MITWA AI LLP.

My primary research and development focus is centered around Large Language Models (LLMs) and Small Language Models (SLMs), especially in the context of autonomous evaluation, prompt optimization, and efficient deployment. I have expertise in designing and building agentic pipelines for tasks like autonomous benchmarking, data synthesis, and exploratory data analysis.

My work emphasizes developing GPU-free and offline-optimized AI systems for low-resource environments, evident in my thesis on Facial-Based Multimodal Body Vitals Detection System and ongoing research on Autonomous SLM-Powered Medical RAG Agent and Synthetic MRI/CT Reconstruction. I also have experience in computer vision, signal processing, and optimization algorithms.

Key Skills include: Python, C++, SQL, Microsoft Azure, AWS, Oracle Cloud Infrastructure, and developing models utilizing deep learning and machine learning techniques. I have also contributed to several peer-reviewed publications in areas like AI-driven contactless vital monitoring, CKD diagnosis, and multi-stage renal cell carcinoma classification.

Expericence

Jul 2025 - Present Began role as Research Lead & AI Developer at MITWA AI LLP, Mumbai. Key achievements include building a fully autonomous dynamic prompt-optimization pipeline and a multi-turn benchmarking framework for Mitwa LLM.

May - Jun 2025 Completed Internship at Prodware Solutions Pvt Ltd, Thane, where an automated invoice processing system with a generative AI chatbot was built.

Jan - Apr 2025 Completed Internship at Data Science Wizards Pvt Ltd, Thane. Developed a deep learning-based biometric identification model for cattle muzzle patterns, achieving 98% verification accuracy and reducing fraudulent insurance claims by over 67%.

May - Aug 2024 Completed Internship at Fermion Infotech Pvt Ltd, Navi Mumbai. Engineered and integrated a YOLOv5s object detection and segmentation model for real-time automobile classification and traffic monitoring.

May - Jul 2023 Completed Internship at MOBA Mobile Automation Pvt Ltd, Gujarat. Designed a real-time hybrid machine learning pipeline (LSTM autoencoders + SHAP) on multi-sensor data, improving prediction accuracy for machine productivity.

Education

Narsee Monjee Institute of Management Studies (NMIMS)June 2021 - Aug 2025

Bachelor of Technology in Artificial Intelligence and Data Science

CGPA: 3.31/4.0

Bachelor Thesis: Facial-Based Multimodal Body Vitals Detection System from 20-second Video using AI and Signal Processing without GPU Acceleration - Advisor: Dr. Toral Shah and Dr. Anindita Khade.

Publications

2025

AI Driven Contactless Body Vital Monitoring through 20-second Facial Video Analysis for Universal Access

K. Talwalkar, G.M. Keshari, T. Shah, and A. Khade

In Advancing Healthcare 5.0, Elsevier, 2025 (Abstract accepted, manuscript submitted)

Scalable Neural Network with Multistage Feature Selection for CKD Diagnosis

A. Khade, G. M. Keshari, and K. Nadar

Recent Advances in Electrical & Electronic Engineering (RAEEE), Bentham Science Publishers, 2025 (Accepted, to be published)

A Probabilistic Dual-Stream Deep Learning Framework for Multi-Stage Renal Cell Carcinoma Classification from CT images

A. Khade, and G. M. Keshari

Engineering Letters, International Association of Engineers (IAENG), 2025 (2nd Revision submitted)

Machine Learning and Big Data in Smart Grid Renewable Energy Systems: Trends and Challenges

G. M. Keshari, S. N. Ansari, and K. Navale

International Journal of Power Electronics Controllers and Converters, Volume 11, Issue 1, 2025

A Novel Technique for Chronic Kidney Disease Prediction using Glowworm Swarm Algorithm with Adaptive Neuro-Fuzzy Inference System

A. Khade, D. Powle, and G. M. Keshari

Journal of Electronics, Electromedical Engineering, and Medical Informatics Scopus Indexed - Vol. 7, Issue 2, 2025

2024

Hybrid NLP Model for Multilingual Sentiment and Emotion Analysis in Poetry

K. Patel, G. M. Keshari, D. Powle, S. N. Ansari, T. Chavan, and A. Khade

2nd IEEE International Conference on Artificial Intelligence and Quantum Computation-Based Sensor Applications (ICAIQSA). 20th-21st December 2024.

Improving Plastic Bottle Waste Management System of India using RVMs

D. Powle, K. Patel, G. M. Keshari, J. Kothari, and A. Kasar

Journal of Energy Environment & Carbon Credits, Volume 14, Issue 3 2024.