Shourya Gupta

LinkedIn CodingNinja | CSES | LeetCode

Chennai, IN.

About

I’m Shourya Gupta, a computer science student and software developer passionate about full-stack development, machine learning, and AI. I enjoy solving challenging problems and working in fast-paced environments while constantly exploring new technologies.

Work

Partyhub AI

|

Full Stack AI Intern

Highlights

Developed an RAG (Retrieval-Augmented Generation) pipeline for FAQs, enhancing response accuracy and retrieval efficiency.

Built and integrated FastAPI-based APIs to deploy the RAG system for seamless interaction.

Automating business intelligence workflows using the CrewAI framework, leveraging multi-agent systems to optimize data processing and insights generation.

Developed full-stack dashboards using React, FastAPI and MongoDB for managing data and user interactions.

CDAC, Pune

|

Software Intern

Highlights

Developed a Table Extraction Model: Extracted tables from documents and converted them into Excel sheets, enhancing data accessibility by 30%.

Utilized DETR with PyTesseract OCR: Achieved 90% table detection accuracy and converted image data into editable formats within 5 seconds.

Education

Vellore Institute Of Technology, Chennai (VIT Chennai)

Bachelor of Technology Computer Science And Engineering Grade: 8.72 CGPA

Awards

Accomplished with a record of 700+ challenges conquered on leading platforms like LeetCode and Coding Ninja

Attained the prestigious position of Software Lead at Team Aviators International (UAV Special Team).

Earned certifications in both Machine Learning and Full Stack MERN.

Skills

Languages

C++, Python, Java, JavaScript, TypeScript.

Frontend

HTML, CSS, React.js, React Native.

Backend

Node.js, Express.js, GoLang.

Generative AI & Frameworks

Lang Chain, CrewAI.

Dev Tools

Git, GitHub, Docker, Postman, JWT, VS Code.

Databases

MySQL, MongoDB, Postgres.

Firmware

Ardupilot.

Projects

AgriSeg

Summary

Developed a Weed-Crop Segmentation Model: Utilized ResNet-50 as the backbone for U-Net to accurately segment weeds and crops in UAV aerial images, achieving an IoU score of 0.66 and a Dice coefficient of 0.79. Achieved High Pixel Accuracy: The model demonstrated a pixel accuracy of 98%, ensuring precise differentiation between weeds and crops. Processed UAV Aerial Images: Handled extensive image preprocessing and augmentation to enhance model training and validation. Integrated with Agricultural Systems: Collaborated with cross-functional teams to integrate the model into existing systems.

RoomConnect

Summary

Developed a Roommate Listing Platform: Built RoomConnect, a web app that connects individuals seeking roommates to share rent, allowing users to post and browse listings based on specified criteria like location, budget, and amenities. Implemented Secure Authentication: Integrated JWT-based authentication and Google OAuth 2.0 for secure user login. Optimized User Experience: Leveraged Redux Toolkit for state management to streamline data flow and improve component performance, resulting in a smoother user experience and reduced page load times by 30%. Real-time Notifications: Implemented a notification system that alerts users of new roommate listings matching their preferences.