Antony Muchai

Key Skills

Software

Appen

Remotasks

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Top Label Types

Bounding Box

Freelancer Overview

I am an experienced AI Data Annotator and Evaluation Specialist with more than two years of experience supporting the development and training of artificial intelligence systems and chatbots. My work focuses on multimodal data annotation, including text, image, video, and audio datasets used to train and evaluate machine learning models. I have hands on experience working with professional annotation tools such as CVAT, Labelbox, Dataloop, SuperAnnotate, and Encord for computer vision and data labeling tasks. In my role, I perform detailed image and video annotation by labeling objects, scenes, and actions with high accuracy. I also work with text datasets for classification, intent detection, and entity tagging, and review audio data through transcription and speech to text quality checks. I support reinforcement learning from human feedback workflows by evaluating, ranking, and providing structured feedback on AI generated responses to improve model behavior, safety, and overall performance. My experience also includes dataset quality assurance where I review labeled data, identify inconsistencies, correct annotation errors, and ensure datasets meet the standards required for machine learning training pipelines. I have worked on AI evaluation projects through platforms such as Mercor and Stellar in remote and collaborative environments. Alongside my annotation experience, I am currently pursuing a Data Science degree at the University of Nairobi, where I am developing skills in data analysis, statistics, machine learning fundamentals, and data management. This academic background strengthens my analytical thinking, attention to detail, and ability to work effectively with complex datasets. I am highly adaptable to new tools and workflows, communicate clearly with distributed teams, and consistently deliver high quality work within tight deadlines.

Labeling Experience

Computer Vision Image and Video Annotation for Autonomous Driving Systems

ImageBounding Box

I worked on a computer vision data labeling project focused on supporting the development of autonomous driving systems. The project involved annotating large datasets of street scenes captured from vehicle mounted cameras to help train machine learning models used in object detection and scene understanding. My responsibilities included labeling vehicles, pedestrians, traffic signs, road markings, and other relevant objects using bounding boxes and polygon segmentation techniques. I also performed frame by frame video annotation and object tracking to capture the movement and interaction of objects across video sequences. The work required careful attention to detail to ensure accurate spatial positioning and consistent labeling across thousands of images and video frames. I used professional annotation tools such as CVAT, Labelbox, and SuperAnnotate to complete annotation tasks efficiently while following strict labeling guidelines provided by the project team. I also participated in dataset quality assurance processes by reviewing annotations, identifying inconsistencies, and correcting errors to maintain high quality training data for computer vision models.

I worked on a computer vision data labeling project focused on supporting the development of autonomous driving systems. The project involved annotating large datasets of street scenes captured from vehicle mounted cameras to help train machine learning models used in object detection and scene understanding. My responsibilities included labeling vehicles, pedestrians, traffic signs, road markings, and other relevant objects using bounding boxes and polygon segmentation techniques. I also performed frame by frame video annotation and object tracking to capture the movement and interaction of objects across video sequences. The work required careful attention to detail to ensure accurate spatial positioning and consistent labeling across thousands of images and video frames. I used professional annotation tools such as CVAT, Labelbox, and SuperAnnotate to complete annotation tasks efficiently while following strict labeling guidelines provided by the project team. I also participated in dataset quality assurance processes by reviewing annotations, identifying inconsistencies, and correcting errors to maintain high quality training data for computer vision models.

2024 - 2024

Multimodal AI Data Labeling and Model Evaluation Project

VideoEntity Ner Classification

I worked as an AI Data Annotator and Evaluation Specialist on a large scale multimodal data labeling project supporting the training and evaluation of machine learning and AI models. The project involved labeling and reviewing datasets across multiple formats including images, videos, text, and audio. For computer vision tasks, I annotated objects, scenes, and actions in images and videos using bounding boxes, polygon segmentation, and object tracking techniques. I used professional annotation tools such as Labelbox, CVAT, Dataloop, SuperAnnotate, and Encord to ensure high precision and consistency across thousands of data samples. For natural language processing tasks, I performed text classification, intent labeling, sentiment tagging, and entity recognition to help train AI systems to better understand user queries and conversational context. I also contributed to Reinforcement Learning from Human Feedback workflows by evaluating, ranking, and providing structured feedback on AI generated responses to improve model accuracy and safety. Additionally, I conducted audio transcription and speech to text validation, ensuring that audio datasets met quality standards for training speech recognition systems. Throughout the project, I followed strict quality assurance procedures including guideline compliance, dataset consistency checks, and peer review processes to maintain high quality labeled datasets used in machine learning pipelines.

I worked as an AI Data Annotator and Evaluation Specialist on a large scale multimodal data labeling project supporting the training and evaluation of machine learning and AI models. The project involved labeling and reviewing datasets across multiple formats including images, videos, text, and audio. For computer vision tasks, I annotated objects, scenes, and actions in images and videos using bounding boxes, polygon segmentation, and object tracking techniques. I used professional annotation tools such as Labelbox, CVAT, Dataloop, SuperAnnotate, and Encord to ensure high precision and consistency across thousands of data samples. For natural language processing tasks, I performed text classification, intent labeling, sentiment tagging, and entity recognition to help train AI systems to better understand user queries and conversational context. I also contributed to Reinforcement Learning from Human Feedback workflows by evaluating, ranking, and providing structured feedback on AI generated responses to improve model accuracy and safety. Additionally, I conducted audio transcription and speech to text validation, ensuring that audio datasets met quality standards for training speech recognition systems. Throughout the project, I followed strict quality assurance procedures including guideline compliance, dataset consistency checks, and peer review processes to maintain high quality labeled datasets used in machine learning pipelines.

2024 - 2024

AI Data annotation

Scale AIVideoSegmentation

I worked on annotating autonomous driving datasets, mainly using CVAT, where I labeled and tracked objects like vehicles, pedestrians, cyclists, traffic signs, and lane markings across video frames. My work involved drawing bounding boxes, polygons, and sometimes segmentation masks to accurately represent objects in real-world road scenes. I handled frame-by-frame video annotation, making sure objects were consistently tracked as they moved, even in complex situations like occlusion, fast motion, or crowded environments. I used features like interpolation and object tracking in CVAT to improve efficiency while maintaining high accuracy. Alongside annotation, I also checked image and video quality—flagging issues like blur, poor lighting, or partially visible objects. I followed detailed guidelines to make consistent decisions and avoided guessing when cases were unclear. I also did quality checks on my work and sometimes reviewed other annotations to ensure everything met the requir

I worked on annotating autonomous driving datasets, mainly using CVAT, where I labeled and tracked objects like vehicles, pedestrians, cyclists, traffic signs, and lane markings across video frames. My work involved drawing bounding boxes, polygons, and sometimes segmentation masks to accurately represent objects in real-world road scenes. I handled frame-by-frame video annotation, making sure objects were consistently tracked as they moved, even in complex situations like occlusion, fast motion, or crowded environments. I used features like interpolation and object tracking in CVAT to improve efficiency while maintaining high accuracy. Alongside annotation, I also checked image and video quality—flagging issues like blur, poor lighting, or partially visible objects. I followed detailed guidelines to make consistent decisions and avoided guessing when cases were unclear. I also did quality checks on my work and sometimes reviewed other annotations to ensure everything met the requir

2024 - 2024

Education

Work History