Java Backend Developer | Full-Stack Software Engineer
B.Tech in Electronics & Communication Engineering — Banasthali Vidyapith
Machine Learning & Computer Vision Research Exposure (NIT Kurukshetra Internship & Final Year Project)
I am a software engineer focused on building robust, secure backend systems in Java & Spring Boot and dynamic full-stack applications with React.js. I emphasize writing clean, testable code, enforcing strong authentication & authorization boundaries, designing relational database schemas, and applying efficient Data Structures & Algorithms to solve practical engineering problems.
- 🔭 Current Focus: Building secure RESTful APIs, Spring Security authentication architectures, and learning database concurrency & transaction management.
- 🎯 Target Roles: Java Backend Developer, Full-Stack Developer, Software Development Engineer (SDE).
- 💡 Core CS: Data Structures & Algorithms, Object-Oriented Programming (OOP), Database Management Systems (DBMS), Operating Systems, Computer Networks.
| Area | Technologies |
|---|---|
| Backend & APIs | Java 21, Spring Boot (3.x), Spring Security 6, Spring Data JPA, Hibernate, RESTful APIs, JWT, BCrypt, Jakarta Validation |
| Frontend | React.js, Redux Toolkit, JavaScript (ES6+), HTML5, CSS3, Tailwind CSS, Vite |
| Databases & ORM | PostgreSQL, MySQL, H2 Database, Hibernate ORM, Flyway Database Migrations |
| Tools & Environments | Git, GitHub, Maven, Docker, Postman, Linux / Bash, MVC Architecture |
| ML & Computer Vision | Python 3.10+, FastAPI, OpenCV, MobileNetV2 / TensorFlow, Librosa, Scikit-Learn |
1. CampusShare — Peer-to-Peer Campus Marketplace (Flagship Project)
Full-stack campus utility sharing platform built with Java 21, Spring Boot 3.5, React 19, and PostgreSQL.
- Concurrency & Locking: Implemented pessimistic database row-locking (
SELECT FOR UPDATE) and optimistic versioning to prevent double-booking collisions during concurrent rental requests. - Security & Authorization: Stateless JWT authentication paired with rotating HttpOnly refresh tokens; backend-enforced ownership checks on all listing mutations.
- Payment Architecture: Razorpay order workflow with server-side HMAC-SHA256 signature verification and an idempotent durable webhook inbox for reliable payment settlement.
- Stack: Java 21, Spring Boot 3.5, React 19, PostgreSQL 16, Flyway, Spring Security, Docker.
2. HireVia — Recruitment & Applicant Tracking System
Full-stack job portal and talent acquisition platform with role-based access control and live ATS workflow.
- Security Hardening: Stateless JWT authentication with externalized secret configuration, BCrypt salt hashing, in-memory rate limiting against brute-force attacks, and server-side IDOR protection across jobs and applicant data.
- Secure Messaging & ATS: Candidate-employer communication channel with JWT-derived identity verification and an ATS pipeline (
APPLIED➔SHORTLISTED➔INTERVIEW➔HIRED/REJECTED). - Stack: Java 21, Spring Boot 3.5, React 18, Vite, Redux Toolkit, Tailwind CSS, Spring Security 6.
3. AI Virality Predictor — Video Intelligence & Platform Optimizer (NIT Kurukshetra Research Internship Project)
Multimodal video analytics engine and regression benchmark evaluating video pacing across YouTube Shorts, TikTok, and Instagram Reels.
- Feature Extraction: Extracted optical motion flow (0–3s hook intensity) via OpenCV and acoustic RMS energy peaks via Librosa.
-
Machine Learning Benchmark: Baseline regression comparison on a controlled synthetic multimodal benchmark (
$N=10,000$ ) where Linear Regression proved optimal ($R^2 = 0.8824$ ,$\text{RMSE} = 4.5194$ ) due to the underlying linear target structure. - Project Evolution: Originated as a Flask research prototype during my NIT Kurukshetra internship, later independently expanded into a FastAPI + Next.js 14 platform with automated platform blueprints.
- Stack: Python, FastAPI, Next.js 14, OpenCV, Librosa, Scikit-Learn, Tailwind CSS.
4. FloodGuard AI — Real-Time Flood Hazard & Risk Monitoring (Final Year Computer Vision Project)
AI and Computer Vision prototype developed as a Final Year Project in B.Tech ECE (Banasthali Vidyapith) for flood-scene analysis, estimated surface motion/flow analysis, and community risk alerting.
- Computer Vision Pipeline: Combines a MobileNetV2-based CNN for frame-level flood scene classification with Lucas-Kanade optical flow for relative surface motion tracking and estimated flow analysis.
- Verified Architecture: Saved model contains a MobileNetV2-based frozen backbone (~2.26M non-trainable parameters) with a custom dense binary classification head (~164K trainable parameters), plus a separate reproducible retraining module under
training/. - Telemetry & WebSockets: Low-latency video stream pipeline with MJPEG feed and WebSocket channels for location-scoped emergency telemetry.
- Stack: Python 3.10+, FastAPI, OpenCV, MobileNetV2-based CNN (TensorFlow), SQLite, WebSockets.
- Research Intern — National Institute of Technology (NIT), Kurukshetra
- Researched Computer Vision and Machine Learning algorithms for multimodal feature extraction, optical flow motion analysis, and audio signal processing.
- Built the initial research prototype in Python/Flask that laid the algorithmic foundation for the AI Virality Predictor.
- Actively practicing Data Structures & Algorithms (Arrays, Hash Tables, Trees, Graphs, Dynamic Programming, Two Pointers).
- Object-Oriented Design (SOLID principles, Factory, Strategy, and Singleton design patterns).
- Relational database schema design, indexing, transactions (ACID properties), and SQL optimization.
- LinkedIn: linkedin.com/in/angel-mishra-992474345
- GitHub: github.com/123angmish