Maxwell Sabwa

Key Skills

Software

CVAT

Doccano

Label Studio

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Video

Top Task Types

Bounding Box

Classification

Computer Programming Coding

Data Collection

Object Detection

Freelancer Overview

With over four years of hands-on experience as a data annotator and AI training specialist, I’ve worked extensively across diverse datasets including text, images, audio, and video. I’ve contributed to the development and refinement of machine learning models by labeling complex data for use in natural language processing, computer vision, and speech recognition tasks. My experience includes working with tools like Label Studio, CVAT, and Doccano, and ensuring high-quality, consistent annotations through rigorous QA processes and version control practices. What sets me apart is my attention to detail, deep understanding of model training needs, and the ability to adapt guidelines for evolving project requirements. I’ve supported teams in scaling annotation pipelines, training new labelers, and designing annotation workflows that align with model objectives. I’ve contributed to projects in sentiment analysis, named entity recognition, image segmentation, and multi-language classification, always prioritizing data integrity and alignment with model goals. My background bridges technical execution with contextual understanding, helping AI systems learn from clean, accurate, and relevant data.

Labeling Experience

3D Object Detection for Autonomous Vehicles

Scale AI3D SensorBounding BoxPolygon

I contributed to a large-scale autonomous vehicle dataset aimed at improving object detection and spatial awareness for self-driving systems. This project required high-precision annotation of urban and highway environments, including labeling vehicles, pedestrians, road signs, lane markings, traffic signals, and obstacles. I used 2D bounding boxes, 3D cuboids, and pixel-level segmentation to ensure detailed scene understanding. The dataset included multi-camera footage and LiDAR sensor data, spanning over 100,000 frames with multi-object tracking across sequences.

I contributed to a large-scale autonomous vehicle dataset aimed at improving object detection and spatial awareness for self-driving systems. This project required high-precision annotation of urban and highway environments, including labeling vehicles, pedestrians, road signs, lane markings, traffic signals, and obstacles. I used 2D bounding boxes, 3D cuboids, and pixel-level segmentation to ensure detailed scene understanding. The dataset included multi-camera footage and LiDAR sensor data, spanning over 100,000 frames with multi-object tracking across sequences.

2023 - 2024

AI Classification for Healthcare

Label StudioTextEntity Ner ClassificationClassification

I annotated over 25,000 medical text entries to support the training of a virtual assistant capable of understanding patient queries and providing accurate, context-aware responses. Tasks included Named Entity Recognition (NER) for symptoms, conditions, medications, and procedures, as well as multi-label classification of patient intents. I also contributed to the creation of high-quality question-answer pairs and natural-sounding dialogue prompts for supervised fine-tuning.

I annotated over 25,000 medical text entries to support the training of a virtual assistant capable of understanding patient queries and providing accurate, context-aware responses. Tasks included Named Entity Recognition (NER) for symptoms, conditions, medications, and procedures, as well as multi-label classification of patient intents. I also contributed to the creation of high-quality question-answer pairs and natural-sounding dialogue prompts for supervised fine-tuning.

2021 - 2023

Education

Work History