Hello There. I'm an AI enthusiast pursuing a Master's in Artificial Intelligence at Northeastern University, Boston. With a B.Tech in Computer Science and Engineering (specializing in Data Science and ML) from Manipal University Jaipur, I’ve worked with multiple organizations on diverse projects, honing my skills in AI/ML, MERN stack, and software engineering methodologies.
Beyond academics, I actively participate in hackathons, social clubs and events, always seeking opportunities to innovate, collaborate, and grow. Passionate about coding and automation, I strive to bridge the gap between theory and real-world impact.
Master of Science in Artificial Intelligence (Khoury College of Computer Sciences)
Bachelor of Technology in Computer Science and Engineering
Earth ArXiv
Indian Space Research Organisation (ISRO) - SAC
Nov, 2024
This study employs deep learning to estimate Land Surface Temperature and Emissivity from Landsat imagery (2018–2023). Using the Single Channel Method and an NDVI-based approach, Pix2Pix models achieve high accuracy, even on unseen data.
Read Full PaperIRJMETS
Aug, 2023
This proposal enhances smart traffic lights using LSTM and GRU for traffic flow prediction, outperforming traditional models. Drone-captured images provide critical traffic features, showcasing deep learning's potential in urban mobility.
Read Full PaperResearchGate
National Institute of Technology, Trichy
May, 2023
OTPLM combines NLP, image analysis, and a multi-agent system for plant identification using hybrid semantics. A medicinal herb ontology enhances data, ensuring accurate infographic suggestions.
Read Full PaperResearchGate
National Institute of Technology, Trichy
June, 2023
This paper proposes a specialized document recommendation framework for horticulture, using WikiData ontology, dynamic knowledge tags, and robust learning methods for semantic reasoning.
Read Full PaperIJIRSET
Aug, 2024
Our WasteSegNet, a custom CNN, achieves 87.5% accuracy in automated waste segregation into Organic and Recyclable categories, optimizing waste sorting and promoting sustainable smart city waste management.
Read Full PaperIJPREMS
Sept, 2023
This paper explores subword tokenization in enhancing NLP models like ELMo and BERT, improving handling of OOV words and capturing fine-grained patterns for better contextual understanding.
Read Full Paper
Conducted research in Signal and Image Processing, developing Land Surface Temperature (LST) and Emissivity models using satellite imagery from diverse global regions under varying weather conditions. Leveraged GANs, particularly Pix2Pix, for automated atmospheric predictions, achieving high performance on unseen data.
Developed a Mental Fitness Tracker using machine learning techniques to analyze mental health indicators across various countries and users. Applied data preprocessing on 6,840 records and compared ML models, achieving high prediction accuracy to provide personalized mental health insights.
Annotated 2,000+ drone images from 100 key Surat junctions for a Government project, classifying vehicles into 7 categories. Analyzed traffic patterns to provide insights for transportation infrastructure planning. Delivered high-quality annotated data, split 70-20-10, to the machine learning team for model training and further analysis.
Gained hands-on experience in data analysis, modeling, and visualization while contributing to real-world projects under the excellent mentorship of the Celebal Team, enhancing proficiency in data-driven decision-making with skills in data visualization, Python, discrete mathematics, data science, and machine learning.
Developed Salesforce applications using Apex and LWC, gaining expertise in automation, security, and workflow enhancements. Set up VS Code and CLI for efficient development, collaborated on real-world projects, implemented APIs, debugged code, and optimized processes in an agile environment.
Identified growth opportunities and conducted research to support decision-making, contributing to business growth. Developed skills in corporate communications, Google Analytics, collaboration solutions, strategizing growth, product management, and team building.
I harnessed data-driven insights to support critical decision-making processes, bridging the gap between data and business strategies. Contributing to 3 impactful projects and fostering a data-driven culture.
.jpg)
Assisted in developing and evaluating AI solutions for strategic business use cases, supporting model design, experimentation, and deployment workflows within the hub.
Represented Adobe on campus by engaging students through events and workshops, promoting Adobe tools, and building community awareness around creative and digital skills.
Built and refined quantitative finance models leveraging AI techniques to support trading strategies and risk insights, collaborating closely with advisory teams.
Completed the Forward program focused on problem-solving, structured communication, and digital collaboration, applying the curriculum to practical project scenarios.
Led ML initiatives across collaborative projects, guiding teams on data pipelines, model development, and deployment to deliver impact-driven solutions for partner organizations.
Built at Northeastern University’s Innovaite 2025 Hackathon, SmartCart uses AI to generate ingredient lists, compare vendor prices, and optimize delivery. Developed with Python, Django, PostgreSQL, and React, it enhances grocery shopping efficiency while supporting local businesses through automation and smart meal planning.
GitHub
TripTailor creates customized itineraries based on your preferences, budget, and duration. It integrates weather forecasts, transit options, and optimized recommendations for hotels, restaurants, and attractions, simplifying travel planning. Built at the Social Winter of Code Season 5 Hackathon! 🏆
GitHub
This project automates soccer highlight reel creation using AI, identifying key moments like goals and fouls with CNNs and GRUs. Using the SoccerNet dataset, it detects events, trims videos, and generates highlights, allowing fans and analysts to focus on crucial moments without watching the entire match.
GitHub
This project applies advanced reinforcement learning (TD3, Options Critic) to autonomous car parking in a realistic simulation with vehicle dynamics and collision avoidance. Our RL agent learns efficiently, outperforming basic RL methods. This work advances autonomous navigation, demonstrating RL’s potential in real-world applications and autonomous vehicle research.
GitHub
This project develops an ensemble deep learning approach for leukemia subtype classification, combining CNN, SqueezeNet, and Inception V3. Individual models achieved up to 84.1% accuracy, while the ensemble improved it to 86.1%, enhancing diagnostic precision. This demonstrates ensemble learning’s potential in medical imaging and clinical decision support.
GitHub
Developed a Stack Overflow clone with OCR integration using API Ninjas, enabling text extraction from images for seamless interaction. Features include user authentication, question management, interactive responses, voting, tagging, and customizable profiles. Built with React.js, Redux, Express.js, JWT authentication, and a responsive design for an enhanced user experience.
GitHub
'Journey to the Beyond' is a thrilling, text-based choose-your-own-adventure game developed by me and two others in 12th grade, entirely in C++ with immersive sound effects. Set in 2082, players must navigate an alien planet, using logic and strategy to escape. The game challenges players to think creatively, offering an engaging, graphics-free experience that keeps them on edge with its intense and rewarding gameplay.
GitHubBuilt a competitive Texas Hold’em bot in 30 minutes using game theory, probability, and adaptive strategies to assess hand strength, calculate odds, and vary betting for strong, unpredictable play against rival teams.
GitHub
Hackathon project on Azure that delivers one-click thread summaries, context-aware follow-up chat, voice interaction, and persistent query history to speed up exploring and revisiting technical discussions.
GitHub
RAG system for 10-K analysis with hybrid FAISS-BM25 retrieval, MMR + cross-encoder reranking, multi-strategy chunking, and LLM integration (Mistral, Gemini 2.0) in a Streamlit app for PDF processing and dashboards.
GitHub
