Hi, I am Arnav Sharma.

Vehant Technologies Private Limited

Developing AI/ML-based solutions using audio and video processing for physical security and surveillance applications in collaboration with IIT Jodhpur and Vehant Technologies, involving research on state-of-the-art methods in computer vision, audio analysis, and deep learning, along with hands-on experience in training and fine-tuning models for real-world deployment

• Multimodal Deep Learning
• Vision Language Models
• Transformers
• PyTorch
• Scikit-learn
• Python

upGrad

Got hands-on experience in Java, Spring Boot, NextJS, and ReactJS. Contributed to front-end development by building a reusable design system of components using the Storybook library, ensuring consistency across multiple websites. Additionally, implemented RudderStack event tracking across applications to enhance product analytics and user insights.

• NextJS
• Spring Boot
• Java
• ReactJS
• RudderStack

SPARK Indian Institute of Technology Roorkee

Developed a comprehensive web-based multi-hazard flood information system to store and disseminate information on vulnerable, physical and socio-economic indicators. Incorporated different python libraries like GDAL(Geo-spatial Data abstraction library), Folium, Numpy, Pillow etc..

• Django
• JavaScript
• AJAX
• HTML5
• CSS3

Isnartech pvt. ltd.

Worked on the project School Digitalization System in the AI department by developing the project Web Application.. My Roles and responsibilities included developing the back-end of the web application, developing the front-end of the web application and maintaining the database. Incorporated frameworks like Django REST framework to create RESTful APIs.

• Django
• JavaScript
• MySQL
• HTML5
• CSS3

ProctorAI

Proctor AI is an intelligent online proctoring system designed to ensure academic integrity during remote examinations. Built on a microservices architecture, it uses Kafka to process real-time camera feeds, automatically flagging suspicious behavior to ensure exam integrity.

• Apache Kafka
• Docker
• Computer Vision
• Django
• NextJS

Multilingual Speech-to-Text and Speech-to-Speech Content Summarization

Built a multilingual system for transcription, summarization, and speech synthesis in 8 Indian languages (including low-resource ones) using fine-tuned Whisper ASR, mT5, and Indic Parler-TTS, enhanced with synthetic datasets for improved multilingual performance.

• Python
• Transformers
• NLP
• ASR
• TTS

Ultrasound Image Segmentation in Human Carotid Artery

B.Tech. Project

Developed and evaluated machine learning and deep learning models (U-Net, U-Net 2p, SAM, Iter-Net) for carotid artery ultrasound image segmentation to enhance plaque detection for potential stroke prevention.

• Python
• PyTorch
• Deep Learning
• Computer Vision

Jagatsinghpur Flood Management

A first-of-its-kind web-based flood information system to facilitate efficient Flood Risk Management over the Jagatsinghpur district in Odisha, infamous for its high susceptibility to multiple hazards from river flooding, coastal flooding and cyclonic impacts.

• Django
• Python
• AJAX
• JavaScript
• CSS3

The Cyclone Watch

Contributed to developing the back-end of an Android application that uses INSAT-3D IR images to observe and learn data, providing warning signals for cyclones in tropical regions.Developed RESTful APIs to send data predictions from the CNN model to frontend of the application.

• Django
• REST
• Python

Tech Nexus Website

Contributed to developing the front-end of the website for Technical Nexus 2022 (An International Level Tech Fest of NIT Hamirpur).The Fest had more than 1000 participations from all over the world in different Events, Competitions and Guest lectures.

• ReactJS
• CSS
• JavaScript