Lokesh M.

Key Skills

Software

CVAT

LabelImg

Roboflow

Top Subject Matter

Industrial PPE safety detection (computer vision)

Structured dataset labeling (AI training data preparation)

Industrial safety PPE object detection dataset

Top Data Types

Image

Top Task Types

Bounding Box

Classification

Object Detection

Freelancer Overview

Data Labeling and Annotation Specialist (Freelance AI training data projects, remote). Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include CVAT, LabelImg, and Roboflow. Education includes Bachelor of Technology, Lovely Professional University (2022) and Intermediate, Sri Venkateshwara Junior College (2022). AI-training focus includes data types such as Image and labeling workflows including Bounding Box, Classification, and Object Detection.

Labeling Experience

Data Labeling and Annotation Specialist (Freelance AI training data projects, remote)

CVATImageBounding Box

Annotated 5,000+ computer vision images for PPE safety detection by creating bounding boxes and polygon masks while keeping label accuracy above 98% across deliverables. Built multi-class labeled datasets for PPE detection, including person detection, safety equipment identification, and NO-class absence flagging. Performed QA review to correct mislabeled bounding boxes and maintain dataset integrity for downstream training. • Data labeling for 8 PPE classes including Gloves, Hardhat, Vest-Overall, Person, and negative NO classes • Dataset delivery throughput of 1,000–2,000 annotated images per week • QA and label validation to reduce errors prior to model training • Automated export workflows using common CV formats (e.g., YOLO/COCO/Pascal VOC)

2022 - Present

Multi-Class AI Annotation and Evaluation Pipeline (remote)

Image

Annotated multi-class datasets and evaluated AI output quality at scale using Python-based workflows. Applied QA-oriented dataset and evaluation processes to improve model accuracy by 40%. Supported multi-class AI training by verifying label quality and measuring performance during iterative refinement. • Multi-class dataset annotation for AI training • Output quality evaluation at scale • QA workflows to drive accuracy improvements • Metrics-driven refinement based on evaluation results

2025 - 2026

Data Structures Trainee (CipherSchools—Summer Industry Training, remote)

ImageClassification

Labeled and reviewed structured datasets during remote summer industry training to build consistent data quality practices. Applied QA-oriented review habits to ensure labeling correctness across provided exercises. Focused on disciplined annotation workflows suited for AI training data preparation. • Remote dataset labeling and review exercises • Emphasis on data quality and annotation discipline • QA/checking of structured labels • Supporting AI-ready dataset preparation

2024 - 2024

PPESafetyDetectionDataset—8-Class Bounding Box Annotation (remote)

CVATImageObject Detection

Annotated industrial safety images across eight PPE classes using bounding boxes in YOLO format. Achieved 97%+ inter-annotator agreement on 2,000+ image batches. Flagged occluded and ambiguous objects with reviewer notes to reduce downstream model false positives by 25%. • 8-class bounding box annotation for PPE safety detection • YOLO format exports for training-ready datasets • Inter-annotator agreement tracking and consistency improvements • Reviewer notes for occlusions and ambiguities

2023 - 2024

Education

Work History