System Engineer Intern · EdgeVerve Finacle, Bengaluru
B.Tech Information Technology · Hindusthan Institute of Technology, Coimbatore
const magesh = {
role : "Full-Stack Developer & ML Engineer",
internship: "System Engineer @ EdgeVerve Finacle (Bengaluru)",
education : "B.Tech IT — Hindusthan Institute of Technology",
building : ["full-stack web apps", "ML & IoT systems", "disaster early-warning tools"],
stack : ["Angular", "Spring Boot", "Node.js", "Python", "MySQL"],
openTo : ["full-time software engineering roles"],
};
Traditional weather forecasting misses hyperlocal cloudburst triggers. This system deploys ESP32 sensor nodes in the field to monitor soil moisture, rainfall intensity, temperature, and humidity — transmitting over two independent channels simultaneously:
The dual-channel design ensures the system keeps alerting locally even when the network is down — critical for disaster-prone, connectivity-poor regions.
|
Engineered a sensor-embedded smart pillow using ESP32 to continuously monitor pulse, respiration, body temperature, pressure, and sleep posture. Sensor data streams in real time to a Node.js REST API, persisted in MongoDB, and classified by a Random Forest model to predict insomnia risk.
|
Architected and deployed a production-grade restaurant platform featuring real-time order state management, reservation flows, menu filtering, JWT-secured role-based access, and Winston structured logging. Resolved 40+ production-level bugs across the full stack — including in-memory auth caching to eliminate per-request DB queries, RxJS
|
Built a layered Java backend (controller → service → repository) exposing a JWT-secured REST API, consumed by a React frontend. Implements department hierarchies, employee CRUD, and role-based access control — mirroring patterns used in production enterprise HR systems. Demonstrates core Spring Boot competencies directly applicable to backend engineering roles: bean lifecycle, JPA/Hibernate ORM, DTO mapping, exception handling, and stateless JWT auth.
|
Implemented a real-time eye-aspect-ratio (EAR) algorithm on a live webcam feed using OpenCV to detect sustained eye closure. On drowsiness detection, the system triggers an audio alarm via Pygame and dispatches an email alert with the driver's GPS coordinates using smtplib and GeoPy — sub-second end-to-end response.
|
Built and shipped a full MERN stack to-do application with deadline tracking, overdue detection, and persistent task state. Frontend deployed on Netlify, backend on Render — live and publicly accessible. Demonstrates end-to-end ownership: React SPA, Express REST API, MongoDB data layer, CORS configuration, environment-based config, and cloud deployment.
|
| Project | Description | Stack |
|---|---|---|
| Student Attendance Tracker | Face-recognition-based attendance system with alcohol detection via sensor integration | Python · OpenCV · ESP32 · MongoDB |
| Angular ToDoList | Standalone Angular 21 to-do app — live Netlify deploy, no external dependencies | Angular · TypeScript |
| FlappyBird | Fully playable Flappy Bird built in Java Swing — physics, collision detection, live scoring, runs as a standalone .jar |
Java · Swing/AWT |