Frontend
- React
- Vite
- JavaScript
- HTML
- CSS
Open to Work – Full-Stack / Backend Roles
Hello, I’m Varun Gupta
Aspiring Software Developer | Full Stack Enthusiast
I enjoy turning complex problems into clear, reliable software—building scalable apps and learning something new with every project.
- Fresher Experience Level
- 0 Major Projects Built
- 0 DSA Problems Solved
- 0 CGPA
- 0 NCAT Percentile
Who I am
Computer Science Engineering student at P P Savani University with a strong foundation in Data Structures and Algorithms using Java. I’m skilled in object-oriented programming, backend development, and full-stack web development with JavaScript, Node.js, and MongoDB.
I’m passionate about building scalable applications and solving real-world problems with thoughtful design and solid engineering.
Education
BTech, Computer Science Engineering
P P Savani University
Currently learning
Expanding backend depth, cloud deployment, and system design fundamentals.
-
Spring Boot
Building production-grade REST APIs, dependency injection, and enterprise Java patterns.
-
Docker
Containerizing full-stack apps and composing local dev environments with Docker Compose.
-
AWS
Exploring cloud deployment, managed databases, and scalable hosting for side projects.
-
System Design
Studying scalability, caching, load balancing, and designing reliable distributed systems.
What I work with
Backend
- Node.js
- Express.js
- Java
- Python
Database
- PostgreSQL
- MySQL
- MongoDB
AI & Automation
- Prompt Engineering
- Workflow Automation
- LLM Applications
- AI Integrations
Tools
- Git
- GitHub
- VS Code
- Postman
- Docker
Featured work
End-to-end full-stack and AI-powered applications—each with an expandable case study covering problem, solution, and key learnings.
Featured projects
Flagship builds spanning AI applications, analytics platforms, and full-stack web systems.
Featured
Aurora AI
AI health companion
AI-powered healthcare companion that provides intelligent assistance, symptom guidance, and personalized health interactions using modern AI technologies.
Tech stack
View case study
Problem
Users need accessible health guidance without navigating fragmented information or generic search results.
Solution
Built an AI health companion with a React frontend and Express API, combining structured health flows with intelligent assistance.
Features
- Symptom guidance & health interactions
- JWT authentication & user profiles
- PostgreSQL data layer with Prisma
- Responsive Vite + React UI
Tech stack
- React, Vite, Express.js, PostgreSQL, Prisma, AI Integrations
Key learnings
Learned to separate frontend and backend concerns, manage environment variables across deployments, and design API-first features for an MVP health product.
Featured
Creators Analytics Platform
AI-powered creator dashboard
AI-powered analytics dashboard helping creators monitor engagement metrics, audience insights, and content performance.
Tech stack
View case study
Problem
Creators see metrics for YouTube and Instagram but struggle to understand why one video outperforms another—raw analytics lack transcript and content context.
Solution
Built a RAG-based analytics platform that compares two videos, stores transcript embeddings in Qdrant, and answers performance questions via a LangGraph workflow with Gemini.
Features
- YouTube + Instagram Reel comparison
- Metadata & transcript extraction
- Vector search with source citations
- Streaming chat with conversation memory
- Engagement rate calculation
Tech stack
- React, TypeScript, Vite, Express, PostgreSQL
- LangGraph, Gemini, OpenAI Embeddings, Qdrant
- Docker, yt-dlp, FFmpeg
Key learnings
Deployment taught me to manage environment variables across Vercel and Render, plan APIs before UI polish, design PostgreSQL schemas for chat history, and handle production rate limits when extracting YouTube data on shared cloud infrastructure.
Featured
ExamEase
Online examination system
Full-stack examination platform with role-based access for admins and students, secure exam workflows, and real-time progress tracking.
Tech stack
View case study
Problem
Colleges and training institutes need a secure, digital way to conduct exams without paper overhead, manual grading bottlenecks, or weak access control.
Solution
Developed ExamEase—a full-stack online examination system with separate admin and student portals, secure authentication, and structured exam workflows.
Features
- Role-based authentication (admin & student)
- Exam creation and scheduling
- Real-time progress tracking
- Admin dashboard for oversight
- Password reset & OTP flows
Tech stack
- Java, Spring Boot, MongoDB
- HTML, CSS, JavaScript
- REST APIs, Vercel deployment
Key learnings
Strengthened backend design with Spring Boot, modeled exam data in MongoDB, and learned to ship an MVP by prioritizing auth and core exam flows before secondary features.
Featured
CodeSpark
Code learning platform
Interactive coding practice platform with structured lessons, user authentication, and REST API–backed progress tracking.
Tech stack
View case study
Problem
Beginners need structured coding practice with progress tracking—not scattered tutorials without accountability.
Solution
CodeSpark delivers leveled coding lessons with authentication, dashboards, and API-backed progress persistence.
Features
- Beginner to advanced lesson tracks
- User authentication & dashboards
- Progress tracking
- REST API integration
Tech stack
- Java, Spring Boot, MongoDB, HTML, CSS, JavaScript
Key learnings
Practiced designing user-centric learning flows, structuring MongoDB collections for progress data, and testing REST endpoints with Postman before frontend integration.
Other projects
Additional builds across fintech, automation, NGO websites, and developer tooling.
AI automation
Workflow automations built with Python, LLMs, and third-party APIs for outreach and communication.
LinkedIn AutoPost
Automated LinkedIn engagement workflow for sending connection requests, follow-ups, and nurturing leads.
Tech stack
View case study
Problem
Manual LinkedIn outreach is repetitive and hard to personalize at scale during internship and job search cycles.
Solution
Built LinkedIn AutoPost—a Python automation pipeline using LLMs to draft context-aware connection messages, follow-ups, and engagement sequences.
Features
- Automated connection request workflows
- LLM-generated personalized outreach
- Follow-up scheduling & lead nurturing
- API-driven automation steps
Tech stack
- Python, LLMs, Automation APIs, Prompt Engineering
Key learnings
Improved prompt engineering for professional tone, designed reusable automation workflows, and learned to balance efficiency with platform-appropriate outreach limits.
Automated Email Builder
AI-powered workflow that generates and sends personalized email campaigns.
Tech stack
View case study
Problem
Writing and sending personalized outreach emails at scale is time-consuming and inconsistent without automation.
Solution
Built an AI-powered email workflow that generates tailored campaign copy with LLMs and dispatches messages through SMTP APIs.
Features
- LLM-generated personalized email content
- Campaign template workflows
- SMTP API integration for delivery
- Batch outreach automation
Tech stack
- Python, LLMs, SMTP APIs, Prompt Engineering
Key learnings
Learned to structure prompt templates for consistent tone, handle SMTP configuration securely, and design automation pipelines similar to those used in LinkedIn AutoPost and creator analytics outreach experiments.
My journey
From coursework and internships to full-stack and AI projects—working toward a software engineering role.
Started BTech in Computer Science
Began building fundamentals in Java, DSA, and web development at P P Savani University.
Basic java and full-stack Projects
Built basic java and full-stack projects like shopping bill generator, AI ticketing system, etc.
Internship at Pragma Infotech
Web Development Intern—built a Database Management System with PHP, MySQL, and a collaborative team of 10–20 members.
Major full-stack & AI projects
Shipped Aurora AI, Creators Analytics Platform, ExamEase, CodeSpark, and automation workflows with production deployments.
Graduation & next goal
Completing BTech CSE and targeting a Full-Stack or Backend Software Engineer position to build scalable products.
Internships
Worked on designing and implementing a Database Management System, collaborating with team members, handling backend logic, and improving application performance.
Technologies
Highlights & achievements
- Built and deployed 15+ full-stack and AI-powered projects.
- Developed automation workflows for LinkedIn outreach and personalized email campaigns.
- Participated in hackathons, internship challenges, and open innovation programs.
- Solved coding problems on LeetCode and CodeStudio while strengthening DSA concepts.
- Contributed to NGOs by designing and developing websites for NayePankh Foundation and She Can Foundation.
- Delivered production-ready applications using React, Express.js, PostgreSQL, and AI integrations.
Blogs & learnings
Practical notes from building, deploying, and shipping full-stack and AI-powered projects.
What I Learned While Deploying Full-Stack Projects
Deploying React frontends on Vercel and Express backends on Render taught me that production is a different skill from local development.
Environment variables were the biggest lesson—API keys, database URLs, and CORS origins must be configured separately for each platform. I learned to use .env.example files, never commit secrets, and validate configs at server startup.
Deployment challenges included cold starts on free tiers, CORS mismatches between frontend and backend URLs, and build failures when TypeScript paths weren't aligned. Starting with an MVP—auth, one core feature, then polish—helped me ship faster and debug less.
How I Approach Building Major Projects
A repeatable approach: define the problem, sketch data models, plan APIs, build an MVP, then iterate with real user flows.
I start by writing the problem statement in one paragraph, then list 3–5 must-have features for an MVP. Database design comes next—identifying entities, relationships, and indexes before writing UI code.
API planning means defining REST endpoints with request/response shapes early. For ExamEase, separating admin and student routes clarified auth middleware. For analytics platforms, I plan ingestion pipelines before chat UI. This structure reduces rework and keeps scope manageable.
Lessons from Building an AI Creator Analytics Platform
Building a RAG pipeline with LangGraph, Qdrant, and Gemini surfaced hard lessons in embeddings, prompt engineering, and cloud extraction limits.
The pipeline extracts YouTube and Instagram metadata, chunks transcripts, embeds them with OpenAI, and stores vectors in Qdrant. LangGraph orchestrates retrieval → context assembly → Gemini generation with streaming responses and source citations.
Key learnings: vector databases need metadata filters per video; prompt engineering must instruct the model to cite sources; Docker Compose simplifies local Qdrant + PostgreSQL setup. Production yt-dlp on shared hosts hits rate limits—design demos accordingly. Prompt engineering and chunk size tuning significantly affect answer quality.
Communication
English
Fluent
Hindi
Fluent
Gujarati
Intermediate
German
Basic
Let’s connect
Open to Full-Stack and Backend Software Engineer roles, internships, and meaningful collaborations. Reach out anytime.