Melwin Kalayil

Key Skills

Software

AWS SageMaker

CVAT

LabelImg

Prodigy

Roboflow

Scale AI

SuperAnnotate

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Classification

Object Detection

Point Key Point

Segmentation

Freelancer Overview

I am an AI and Computer Vision Specialist with experience in data labeling, model evaluation, and real-time AI system development. My expertise includes bounding box annotation, object detection, and classification, particularly in computer vision applications such as workplace safety monitoring using YOLOv11 and Raspberry Pi. I have hands-on experience in IoT, embedded systems, and cloud computing, allowing me to contribute to AI training across multiple domains. I have worked on projects involving real-time workplace safety detection, utilizing image and video annotation for AI models. Additionally, I have experience in LLM evaluation, text annotation, and AI model feedback, making me proficient in training AI systems across computer vision, IoT, and industrial safety applications. My ability to bridge hardware and AI software solutions enables me to deliver high-quality labeled data for cutting-edge AI models.

Labeling Experience

Real-Time Workplace Safety Detection – Image & Video Annotation

RoboflowVideoBounding BoxSegmentation

This project focused on annotating images and videos to train a YOLOv11-based AI model for workplace safety monitoring. The primary objective was to detect safety compliance and hazards in real-time using Raspberry Pi and computer vision techniques. The annotation tasks included: Bounding box labeling for helmets, vests, and hazardous objects. Classification of workers based on safety compliance. Segmentation to improve object detection accuracy in complex environments. Ensuring high-quality labeled data by following annotation consistency guidelines and performing manual validation checks. This dataset was crucial in enhancing the AI model’s ability to detect and alert safety violations in industrial settings like oil & gas and construction sites.

This project focused on annotating images and videos to train a YOLOv11-based AI model for workplace safety monitoring. The primary objective was to detect safety compliance and hazards in real-time using Raspberry Pi and computer vision techniques. The annotation tasks included: Bounding box labeling for helmets, vests, and hazardous objects. Classification of workers based on safety compliance. Segmentation to improve object detection accuracy in complex environments. Ensuring high-quality labeled data by following annotation consistency guidelines and performing manual validation checks. This dataset was crucial in enhancing the AI model’s ability to detect and alert safety violations in industrial settings like oil & gas and construction sites.

2024

Education

Work History