John Rowe

Key Skills

Software

AWS SageMaker

Argilla

Dataloop

Datasaur

Encord

Google Cloud Vertex AI

Labelbox

Label Studio

Roboflow

SuperAnnotate

Other

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

3D Sensor

Image

Text

Top Task Types

Classification

Entity Ner Classification

Fine Tuning

Prompt Response Writing SFT

RLHF

Freelancer Overview

I have extensive experience developing and managing AI training datasets across NLP, computer vision, and large-scale language model workloads. As a senior data engineer and ML practitioner, I have experience building data pipelines, preparing high-quality training corpora, designing annotation schemas, and implementing validation workflows that ensure clean and stable inputs for downstream model performance. I have firsthand experience with classification tasks, NER pipelines, text generation datasets, supervised fine-tuning (SFT), RLHF-enabled evaluation sets, and rapid engineering workflows for both small and large-scale models. With experience using Label Studio, Labelbox, SuperAnnotate, AWS SageMaker, Vertex AI, and our own proprietary platforms, I can set up annotation workflows, design QA layers, and perform dataset versioning as part of the ML lifecycle. Beyond technical execution, I understand dataset quality, bias mitigation, ontology building, and data governance principles, ensuring the datasets I create align with real-world product requirements and scalable ML systems.

Labeling Experience

Computer Vision Annotation — Segmentation, Detection, and Tracking

SuperannotateVideoPolygonSegmentation

Worked on a computer vision dataset involving detailed object segmentation, polygon-based annotation, and class-level tagging across a mix of high-resolution images and short video clips. Used SuperAnnotate’s polygon, brush, and tracking tools to outline objects with high precision, maintain consistency across frames, and apply standardized class labels. Reviewed and corrected annotations for edge cases, occlusions, and overlapping shapes to meet accuracy thresholds. Also validated sample batches against project guidelines to ensure annotation consistency, proper class usage, and clean mask boundaries. This dataset was used for model training and evaluation in object detection, segmentation, and sequence-based tracking tasks.

Worked on a computer vision dataset involving detailed object segmentation, polygon-based annotation, and class-level tagging across a mix of high-resolution images and short video clips. Used SuperAnnotate’s polygon, brush, and tracking tools to outline objects with high precision, maintain consistency across frames, and apply standardized class labels. Reviewed and corrected annotations for edge cases, occlusions, and overlapping shapes to meet accuracy thresholds. Also validated sample batches against project guidelines to ensure annotation consistency, proper class usage, and clean mask boundaries. This dataset was used for model training and evaluation in object detection, segmentation, and sequence-based tracking tasks.

2023 - 2023

Education

Work History