Full-Stack Developer
MERN • APIs • Computer Vision
I build full-stack web applications using the MERN stack, with a strong focus on backend logic, APIs, and data-driven features. I also explore Python and computer-vision systems where the problem fits.
Who I Am
I'm a final-year IT student at K. J. Somaiya Institute of Technology passionate about building practical software systems, from full-stack web applications to applied machine learning projects.
I enjoy working across the stack using the MERN ecosystem, designing APIs, and experimenting with computer vision models. Through academic and personal projects, I’ve explored backend development, database design, and deploying real-world applications.
I’m continuously learning how to build reliable, scalable systems and improve performance through better architecture and optimization. I’m currently seeking opportunities where I can grow as an engineer while contributing meaningful solutions.
Quick Facts
Projects
Full-stack applications demonstrating my skills in MERN stack development, backend APIs, and system design.
Schedulr – Lecture Scheduling System
Complete MERN application for managing online lecture scheduling with real-time conflict prevention. Features role-based access control (admin/instructor), JWT authentication, and MongoDB-enforced scheduling rules. Built with production-grade error handling and CORS security.
What I Built
- Backend conflict prevention: Prevents instructors from multiple lectures on same date
- JWT authentication with bcrypt password hashing
- Role-based dashboards (admin: create courses/lectures, instructor: view assignments)
- MongoDB schema with proper relationships and validations
- REST APIs for courses, instructors, and lecture management
- Axios interceptors for automatic token attachment and refresh
- Protected routes with ProtectedRoute component
What I Learned
- Why backend-first validation matters more than frontend checks
- Preventing race conditions in concurrent scheduling requests
- Designing scalable cloud deployments (Vercel + Render + MongoDB Atlas)
- Handling token expiration and automatic refresh workflows
PaySecure – Digital Wallet
MERN-based digital wallet system focusing on transaction integrity and security. Implements JWT authentication, MongoDB ACID transactions, and comprehensive audit logging for all wallet operations. Built to handle concurrent transactions safely.
What I Built
- REST APIs for wallet operations (deposit, withdraw, transfer, history)
- JWT-based authentication with secure session management
- MongoDB transactions to ensure balance consistency
- Audit logging system for compliance and traceability
- Rate limiting to prevent abuse (10 requests/minute per user)
- React dashboard for transaction monitoring and history
What I Learned
- Discovered race conditions in balance updates during concurrent requests
- Why database transactions are critical for financial applications
- How to implement proper retry logic for transaction failures
- Importance of audit trails for debugging and compliance
IoT Sensor Data Ingestion Service
Node.js backend service that ingests IoT sensor temperature readings in real-time. Supports REST API ingestion and MQTT broker subscription for distributed sensor networks. Built with MongoDB persistence, automatic timestamp handling, and schema validation for data integrity.
What I Built
- REST API endpoints for sensor data ingestion (POST /api/sensor/ingest)
- GET endpoint to fetch latest sensor reading per device
- MQTT subscriber for real-time IoT data from distributed devices
- MongoDB schema with automatic timestamp and device indexing
- IoT device simulator for testing multi-device scenarios
- Mongoose validation for temperature ranges and device IDs
What I Learned
- Trade-offs between REST APIs and MQTT for IoT data ingestion
- Handling high-frequency sensor data without database bottlenecks
- Topic structure design for MQTT broker subscriptions
- Timestamp synchronization across distributed IoT devices
AI Content Optimizer
Full-stack application that automates content optimization workflows using LLM APIs. Built backend pipeline for web scraping, prompt engineering, and API orchestration. Features React dashboard for side-by-side content comparison.
What I Built
- Node.js backend with web scraping and content processing pipeline
- OpenAI API integration with prompt engineering and response handling
- Redis caching layer to avoid redundant API calls (85% cost savings)
- Retry logic with exponential backoff for rate limits and failures
- MongoDB schema for storing original and optimized content
- React dashboard with real-time comparison view
What I Learned
- How LLM API costs can spiral without proper caching strategies
- Handling rate limits and transient failures in third-party services
- Trade-offs between response quality, latency, and token usage
- Importance of robust error handling in API-dependent workflows
Mumbai Local Train Crowd Coach Estimator
Backend service for crowd density estimation in Mumbai local trains using YOLO-based passenger detection. Built scalable Flask REST APIs for image processing and inference. Team project where I focused on backend architecture and model integration.
What I Built
- Flask REST APIs for image upload and crowd density classification
- YOLO model integration for passenger detection and counting
- Batch processing pipeline to improve throughput
- Stateless API endpoints for horizontal scalability
- Request validation and error handling for corrupted images
- Classification logic (Low/Medium/High crowd density)
What I Learned
- Trade-offs between model accuracy and inference speed
- How batch processing reduces per-image overhead significantly
- Importance of stateless design for ML inference services
- Handling model inference failures gracefully with retries
Student Engagement Detection System
Real-time computer vision system to monitor student engagement during live sessions. Uses MediaPipe for face detection and attention tracking, with WebSocket-based streaming to deliver real-time analytics to instructor dashboard.
What I Built
- Python backend with MediaPipe face detection integration
- Attention tracking using eye gaze and head pose analysis
- Confusion detection through temporal facial behavior patterns
- WebSocket server for real-time data streaming
- React dashboard with live engagement visualization
- Session management and connection lifecycle handling
What I Learned
- Challenges of maintaining stable FPS on CPU-only systems
- How lighting and camera angles affect computer vision accuracy
- Importance of temporal smoothing to reduce false positives
- Handling partial detection failures without breaking the system
Skills & Technologies
Backend & APIs
Frontend
Databases
AI/ML
Languages
Tools & Platforms
Experience & Leadership
Professional Experience
Search Engine Optimization Intern
Worked on technical SEO optimization, analyzing website structure, metadata, and performance metrics.
- Analyzed keyword performance, traffic patterns, and content optimization
- Collaborated with developers to improve site indexing and visibility
- Worked with content teams on metadata and page structure improvements
Web Developer Intern
Developed full-stack CRUD application using Java backend with MySQL database and responsive Bootstrap frontend.
- Built backend modules using Java, JSP/Servlets, and MySQL
- Created responsive UI with Bootstrap
- Collaborated using Git for version control and code reviews
- Worked on debugging, validation, and database integration
Leadership & Volunteering
Chairperson – IET KJSIT Student Chapter
Led the IET KJSIT student chapter, coordinating technical events and industry engagements.
- Managed team of 50+ member student team across technical, logistics, and marketing verticals
- Owned end-to-end planning and execution of 10+ technical events and workshops
- Coordinated speaker outreach, stakeholder communication, and cross-team delivery
Volunteer Teacher
Taught English and Science to Grade 6–9 students, creating weekly learning modules.
- Conducted weekly remote lessons in English, Science
- Created structured learning modules and exercises
- Mentored students needing additional academic support
- Maintained engagement in a virtual classroom setup
Let's Connect
Open to full-time software engineering roles starting June 2026. Interested in backend development, distributed systems, and ML infrastructure.
GitHub
github.com/eepsita12Location
Mumbai, India
Available: June 2026 (graduation)
Open to internships starting immediately