Industrial & Research Experience

Education

Expected Graduation date: May 2026

University of Michigan - Dearborn 

Masters in Artificial Intelligence with a Specialization in Machine Learning

Courses: Artificial Intelligence, Software Engineering, Data Structures and Algorithm Analysis

December 2020 - May 2024

Indian Institute of Technology (IIT),  Jodhpur

Bachelor of Technology in Artificial Intelligence and Data Science Engineering

Key Courses: Deep Learning, Optimization for Machine Learning, Pattern Recognition And Machine Learning, Advanced Artificial Intelligence, Dependable Artificial Intelligence, Artificial Intelligence, Computer Vision, Data Visualization, Data Engineering, Principles of Computing Systems, Introduction to Blockchain, Security and Applications, Data Structures and Algorithms, Design and Analysis of Algorithms, Maths for Computing

Research Experience

September 2024 - Present

Research Assistant at University of Michigan - Dearborn

I am currently working under Dr.Srijita Das at the Department of Computer and Information Science Department. We are working on building up my profile on Reinforcement Learning and will soon start working on a problem statement, ultimately aiming to achieve a research paper by Summer 2025.

April 2023 - August 2024

Differential Diagnosis of Thyroid Tumors through Information Fusion from Multiphoton Microscopy Images using Fusion 

Autoencoder | Research Paper Publication at ICPR 2024 and Undergraduate Capstone Project 

Co-Authors: Dr.Angshuman Paul, Ms.Tejaswee Anandan

• Thyroid carcinomas are often diagnosed by histopathology, which is widely regarded as the most reliable method. However, alternative imaging modalities may also provide meaningful information about thyroid tumors. 

• Proposed a strategy for the differential diagnosis of thyroid tumors through information fusion from different types of MPM images. 

• Introduced a novel fusion autoencoder (FAE) for this task. The fused information from the FAE is subsequently used by a classifier module for the differential diagnosis of thyroid tumors.

• This method is one of the first approaches to look into the possibility of using MPM images for the diagnosis of thyroid tumors. Extensive experiments demonstrate the superiority of the proposed method compared to a number of state-of-the-art classification techniques.

January 2023 - December 2023

Lightweight DL-Based Drone Image Classification for Surveillance Applications | Research Paper submitted at VLSID 2025 Conference

Co-Authors: Dr.Binod Kumar, Mr.Abhishek Kumar

• Classifying objects in drone imagery is hard because of how different the objects look and how complicated the background is. Traditional classification methods frequently use hand-made characteristics without taking into account the objects' surroundings.

• The goal is to make it easier and more accurate to classify objects in images taken by drones for uses like surveillance, mapping, and inspection.

• Made an attention-based approach to classifying objects, which uses deep learning models to automatically learn distinguishing features and pay attention to relevant parts of an image. Attention mechanisms can be particularly useful for identifying objects that are partially obscured or in difficult-to-see locations, such as cars parked under trees or street lamps in the shadows. 

Industrial Experience

May 2023 - July 2023

Summer Research Internship in collaboration with Swiggy

Computer Science Department at Indian Institute of Technology Jodhpur

Advisor: Dr.Richa Singh

• The software that is being designed will be able to recognize a person’s face through mapping which is based on their nose, eyes, chin, and cheeks, taking in distinctive qualities making each of us different from one another.

• The reliability of this software is being tested for its accuracy and security. Selfies of a person will be matched with their respective ID cards and the model will be trained accordingly to enhance the metrics being taken into account.

• The challenge is to overcome the stability-plasticity dilemma by handling catastrophic forgetting. An additional challenge is imposed if the dataset contains only a few samples of each new class.

• Implemented Constrained FSCIL on the CiFAR-100 dataset and trained D-LORD dataset for facial verification using various models and loss functions to find the best combination so that Constrained FSCIL is performed on the D-LORD dataset using this specific combination of model and hyperparameters.

May 2022 - June 2022

Summer Research Internship

Robert Bosch Centre for Cyber-Physical Systems at Indian Institute of Science, Bangalore

Advisor: Dr.Rajesh Sundaresan, Mr.Jagadish Midthala

• Aimed on creating a marker simulation of people moving from one place to another using different modes of transport in the IIT Jodhpur campus. 

• Leveraged the tool 'Leaflet' to create this marker simulation i.e., to move each mode of transport from one place to another in the campus via the shortest route possible.