Mathematics & Computing
I am Romin Kevadiya, a Mathematics & Computing student at LJ Institute of Engineering & Technology. I build and experiment across software engineering, AI/ML, and computer vision while continuously exploring new technologies.
About Me
I started programming in 2024, initially focused on making things work. As projects grew, my interest shifted to the underlying systems—how data is stored, APIs are structured, and components interact.
This led me to backend development, database design, APIs, and understanding how software systems are structured. My background in Mathematics & Computing also drives my interest in machine learning and computer vision, where I enjoy understanding how the underlying algorithms work and turning them into practical applications.
Beyond code, I actively participate in hackathons and tech communities to learn from others and explore new engineering approaches.
Worked Projects
Pivotly
AI-powered startup idea analysis. Submit a startup idea and generate a structured venture analysis using AI-powered research and analysis.
What Is Pivotly? Pivotly transforms a raw startup idea into a structured venture intelligence report by running it through a Multi-agent AI pipeline that combines web research with structured venture analysis. Each agent is a specialist: a market researcher, a competitive analyst, a contrarian investor, a defensibility expert, and an execution strategist.
Animation Engine
An AI-powered 2D animation tool that helps users create simple motion graphics without manually writing SVG or animation code.
How it works: Instead of hand-authoring timelines, users describe what they want in plain language, and the system generates a validated animation JSON script and plays it instantly.
Civic Fix
A role-based civic issue management platform where citizens can report local problems, government officers handle department-specific complaints, and admins monitor and control the entire system.
What it does: Enables seamless communication between citizens and government officials to track, manage, and resolve civic complaints efficiently through a centralized dashboard.
Beej Vaidya
A crop disease detection system built to explore how a machine-learning model can be developed into a practical application. Trained a TensorFlow model classifying leaf images across Early/Late Blight and Healthy classes, reaching 90.69% test accuracy.
What I explored: Model inference outside notebooks, API-based ML serving, confidence thresholds, validation, and connecting ML models with backend applications.
Madhav Vision
A collection of computer-vision experiments built while learning how image-processing algorithms operate beyond basic library calls. Experimented with object tracking, contour detection, and spatial transformations.
What I explored: Image processing, contours, object tracking, coordinate transformations, and real-time computer vision.
Code Atlas
A project focused on understanding the structure of software repositories. Analyzes source code and represents relationships between files, dependencies, and structural information visually.
What I'm exploring: Static analysis, dependency graphs, architectural hotspots, code relationships, repository visualization.
Research
Personalised Yet Uniform: Algorithmic Curation and the Construction of Generational Identity
This research asks: If recommendation algorithms personalize content, why do large groups still converge on similar trends and behaviors?
The paper examines the link between algorithmic personalization and cultural standardization, exploring feedback loops, generational identity, and platform influence.
Read ResearchTech I Work With
These are technologies I've worked with across projects, coursework, and experiments.
Events & Hackathons
Certificates
Microsoft Azure AI Essentials
Gained foundational knowledge of Artificial Intelligence concepts and their practical implementation using Microsoft Azure AI services.
HackOut at DA-IICT
Competed in HackOut at DA-IICT, rapidly prototyping and developing a creative software solution to solve real-world problems within a tight deadline.
MATLAB Innovation Challenge
Built and presented a working solution during the hackathon.
PDEU Hackathon
Collaborated in a fast-paced environment to build and showcase an innovative technical solution during the PDEU Hackathon.
AWS Workshop at PDEU
Participated in an intensive AWS Workshop at PDEU, exploring cloud computing fundamentals, scalable architecture, and deploying applications on Amazon Web Services.
Java Programming For Beginners
Covered Java fundamentals, object-oriented programming, and basic application development.
UI/UX with ChatGPT
Explored the intersection of generative AI and design, learning how to leverage ChatGPT to enhance UI/UX workflows, ideation, and prototyping.
Get in Touch
Open to learning, building, and collaborating.
Whether you want to discuss a project, collaborate, share ideas, or simply connect, feel free to reach out.