Longmene Bertrand

Key Skills

Software

Scale AI

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Label Types

Bounding Box

Point Key Point

Entity Ner Classification

Segmentation

Tracking

Freelancer Overview

I am an experienced AI training specialist with over three years dedicated to high-quality data annotation and labeling for machine learning projects in computer vision and speech recognition. My background includes extensive hands-on work with image, video, and audio datasets, specializing in tasks such as object detection, semantic segmentation, multi-object tracking, and speech transcription. I am proficient with leading annotation tools including Labelbox, CVAT, Supervisely, Roboflow, V7 Darwin, and Scale AI, and have consistently maintained annotation accuracy above 98% while handling large-scale projects for autonomous driving, retail product classification, and multilingual speech recognition. I excel at collaborating remotely with ML engineers and project leads to refine dataset requirements, implement rigorous quality assurance processes, and deliver precise, structured data in formats like JSON and CSV. My technical skills also include dataset cleaning, guideline development, and basic Python for data handling. I am passionate about improving model performance through meticulous labeling, and I bring strong attention to detail, analytical thinking, and a commitment to meeting strict quality standards on every project.

Labeling Experience

Autonomous Vehicle Object Detection & Multi-Object Tracking Project

Scale AIImageBounding BoxPoint Key Point

Contributed to a large-scale autonomous driving dataset used to train computer vision models for object detection and motion prediction. Responsibilities included: Annotating vehicles, pedestrians, cyclists, traffic lights, and road signs using bounding boxes and segmentation masks. Performing frame-by-frame multi-object tracking across dynamic traffic scenes. Applying cuboid annotations for 3D object localization. Identifying and labeling edge cases such as occlusions, motion blur, and partial visibility. Conducting peer reviews and QA validation to ensure 99% annotation accuracy. Project scale: 60,000+ annotated images and video frames. High-density urban and highway traffic environments. Strict enterprise-level annotation guidelines and quality benchmarks. Quality measures adhered to: Multi-stage QA review cycles. Inter-annotator agreement checks. Continuous feedback integration to improve consistency. Compliance with structured JSON output formatting.

Contributed to a large-scale autonomous driving dataset used to train computer vision models for object detection and motion prediction. Responsibilities included: Annotating vehicles, pedestrians, cyclists, traffic lights, and road signs using bounding boxes and segmentation masks. Performing frame-by-frame multi-object tracking across dynamic traffic scenes. Applying cuboid annotations for 3D object localization. Identifying and labeling edge cases such as occlusions, motion blur, and partial visibility. Conducting peer reviews and QA validation to ensure 99% annotation accuracy. Project scale: 60,000+ annotated images and video frames. High-density urban and highway traffic environments. Strict enterprise-level annotation guidelines and quality benchmarks. Quality measures adhered to: Multi-stage QA review cycles. Inter-annotator agreement checks. Continuous feedback integration to improve consistency. Compliance with structured JSON output formatting.

2023

Video Action Recognition & Object Tracking Project – Appen

CVATVideoBounding BoxPoint Key Point

Contributed to a large-scale video dataset focused on training AI systems for action recognition and real-time object tracking. Key responsibilities included: Annotating human activities such as walking, running, lifting objects, entering/exiting vehicles. Performing frame-by-frame object tracking across dynamic scenes. Drawing precise bounding boxes around moving objects and individuals. Labeling activity categories for supervised learning classification tasks. Handling complex edge cases including occlusions, fast motion, low-light environments, and overlapping objects. Maintaining strict adherence to Appen’s annotation guidelines and quality benchmarks. Project scale: Annotated 25,000+ video frames across multiple datasets. Processed high-resolution surveillance and real-world environment footage. Worked under structured review cycles and performance metrics. Quality measures adhered to: Maintained 98%+ quality score across tasks.

Contributed to a large-scale video dataset focused on training AI systems for action recognition and real-time object tracking. Key responsibilities included: Annotating human activities such as walking, running, lifting objects, entering/exiting vehicles. Performing frame-by-frame object tracking across dynamic scenes. Drawing precise bounding boxes around moving objects and individuals. Labeling activity categories for supervised learning classification tasks. Handling complex edge cases including occlusions, fast motion, low-light environments, and overlapping objects. Maintaining strict adherence to Appen’s annotation guidelines and quality benchmarks. Project scale: Annotated 25,000+ video frames across multiple datasets. Processed high-resolution surveillance and real-world environment footage. Worked under structured review cycles and performance metrics. Quality measures adhered to: Maintained 98%+ quality score across tasks.

2022 - 2022

Education

Work History