Suraj Paul

SURAJ PAUL

ML Lead specializing in Deep Learning, Computer Vision, and Secure IoT through Research-Driven Engineering.

EXPLORE PORTFOLIO

01 Profile_Brief

I am a research-focused BCA (Honours with Research) student and experienced ML Lead. My work sits at the intersection of Artificial Intelligence and hardware security, specializing in building end-to-end autonomous systems from the ground up.

02 Experience_Log

Machine Learning Lead

Vigilantia Praesidium Pvt. Ltd.

OCT 2025 — MAR 2026
Python Deep Learning (CNN) FaceNet MTCNN OpenCV ESP32 Firebase Telegram API ECC
  • Led development of multiple AI-powered systems across domains including smart security, agriculture, and automation.
  • Designed and trained deep learning models for face recognition, object detection, and behaviour analysis.
  • Conducted Research & Development (R&D) to improve model accuracy, optimize performance, and explore innovative AI solutions.
  • Built real-time computer vision pipelines using FaceNet, MTCNN, and OpenCV for practical deployment.
  • Developed and deployed systems like smart home security, attendance monitoring, traffic analysis, and agriculture automation.
  • Integrated IoT devices (ESP32) with AI models for real-world system interaction and control.
  • Implemented secure communication using ECC and AES-GCM encryption in AI-based IoT systems.
  • Integrated APIs (Telegram, Firebase) for real-time alerts, cloud storage, and monitoring.
  • Optimized models for low-latency, real-time inference on laptops and edge devices.
  • Collaborated with teams to convert research ideas into scalable, production-ready solutions.

03 Project_Inventory

Automated Student Attendance Monitoring & Analytics

PythonOpenCVFaceNetMTCNNPandasTelegram API
  • • Implemented an automated attendance system using real-time face recognition to mark student presence with high accuracy.
  • • Built the recognition pipeline using MTCNN for face detection and FaceNet embeddings for face recognition, enabling multi-student detection in a single frame.
  • • Designed a CSV-based storage system for maintaining attendance logs locally, ensuring portability.
  • • Integrated a Telegram bot to provide instant notifications to faculty and daily attendance reports.
  • • Developed a system to handle multiple students simultaneously with averaged embeddings to reduce recognition errors.
  • • Created an analytics module for insights, including daily, weekly, and monthly trends.
  • • Optimized for low-latency processing on standard laptops for cost-effective deployment.

Identity Defenders – AI Security System

ECCAES-GCMESP32FaceNetTelegram API
  • • Real-time face detection with webcam using FaceNet and MTCNN.
  • • Automated alerts sent via Telegram when unknown faces are detected.
  • • Simulated secure ‘lock/unlock’ messaging using ECC + AES-GCM cryptography.
  • • ESP32-controlled solenoid lock for physical access protection.

Road Guard AI – Collision & Drowsiness Detection

Deep LearningYOLOCNNArduinoUltrasonic Sensor
  • • Real-time driver drowsiness detection using eye aspect ratio (EAR) and facial landmark analysis.
  • • Smart collision avoidance using ultrasonic sensor for precise distance measurement.
  • • Instant audio/buzzer alert when driver fatigue or close obstacles are detected.
  • • AI-powered object detection for obstacle identification using YOLO.
  • • Integrated microcontroller-based hardware alert mechanism for real-world deployment.

Smart Agriculture Field System

ESP32FirebaseIoT SensorsPython
  • • Smart irrigation and fertilizer system with automated resource management.
  • • Voice command and WhatsApp alert integration for remote monitoring.
  • • AI-based plant disease detection module currently in R&D phase.