Antony G.

Key Skills

Software

Other

Top Subject Matter

Multimodal AI Systems

AI Training Data & Annotation

Retrieval-Augmented Generation (RAG)

Top Data Types

Image

Text

Medical Dicom

Top Task Types

Segmentation

Bounding Box

Classification

Object Detection

Text Generation

Fine-tuning

Data Collection

Prompt + Response Writing (SFT)

Evaluation/Rating

Freelancer Overview

I am currently pursuing a Master of Engineering at University of Guelph with a focus on AI, machine learning, and computer vision. I have over 2 years of professional experience working on medical imaging, AI training data, and computer vision workflows at Appasamy Associates and Appasamy Ocular Devices. My work has focused on preparing high quality datasets for deep learning models, particularly for medical image segmentation, object detection, and AI-assisted healthcare applications. I have hands-on experience with pixel-level annotation, segmentation workflows, dataset preprocessing, quality validation, and handling medical imaging formats such as DICOM, retinal, OCT, histopathology, and MRI data. I have also worked closely with deep learning teams on model evaluation, dataset consistency, and AI training pipelines. In addition to annotation and dataset preparation, I have developed projects in video anomaly detection, medical image retrieval systems, multimodal embeddings, and segmentation using models such as YOLOv8, MedSAM, DINOv2, and PLIP. My technical skills include Python, OpenCV, PyTorch, and SimpleITK, along with a strong understanding of AI training data pipelines, computer vision systems, and retrieval-based AI workflows including vector embeddings and semantic search.

Labeling Experience

Deep Learning Engineer — Appasamy Associates (medical image dataset preparation for segmentation training)

ImageSegmentation

Built end-to-end medical image data pipelines for segmentation model training, including dataset curation and augmentation to increase training diversity. Trained and evaluated U-Net multi-class segmentation models and used multi-class Dice and class-wise accuracy metrics to iteratively improve performance. Managed preprocessing and dataset expansion to move from small-scale data to a significantly larger labeled dataset suitable for supervised training. • Curated medical imaging datasets and implemented augmentation strategies. • Expanded usable training data from 39 to 2,400+ samples. • Trained/evaluated U-Net segmentation with PyTorch and Dice metrics. • Used class-wise accuracy analysis to guide iterative improvements.

2024 - 2025

R&D Engineer (Computer Vision) — Appasamy Ocular Devices (pixel-level medical image annotation)

Segmentation

Created large-scale pixel-level annotations for structured medical imaging datasets used in computer vision model development. Processed imaging data for preprocessing, normalization, and quality validation to ensure labeling consistency and downstream model reliability. Coordinated with engineering and clinical stakeholders to validate outputs and refine data quality standards. • Produced 500+ pixel-level annotations for medical imaging datasets. • Preprocessed, normalized, and validated structured medical imaging data. • Collaborated to validate outputs and iterate on data quality standards. • Improved downstream model accuracy through higher-quality labeled data.

2023 - 2024

Education

Work History