Peyman P.

Key Skills

Software

No software listed

Top Subject Matter

Network administrator

Cybersecurity

Machine Learning

Top Data Types

Text

Image

Document

Top Task Types

Classification

Text Summarization

Text Generation

Translation/Localization

Relationship

Segmentation

Freelancer Overview

Based on my extensive work in advanced computing, I have deep, end-to-end experience in building, structuring, and optimizing datasets for machine learning. My research in Edge AI heavily relied on creating high-quality data pipelines for Contextually Guided Convolutional Neural Networks (CG-CNNs). This involved managing specialized training datasets, implementing self-supervised contextual guidance, and deploying Expectation-Maximization (EM) refinement to teach models to learn transferable features without relying on massive, traditionally labeled datasets. My experience spans handling structural image constraints, evaluating complex data modalities, and optimizing dataset throughput for real-time deployment on hardware like the NVIDIA Jetson Orin Nano. Additionally, I have highly precise, practical experience in data labeling evaluation, reinforcing data alignment, and instruction-following for computer vision models. I have actively performed detailed data quality analysis on AI-generated images—evaluating complex, split-screen multi-attribute prompts across distinct visual axes such as seasonal contrast, color gradients, and structural dividers to score model accuracy. This specialized background combines the hands-on engineering of machine learning data pipelines with the analytical rigor required to audit, filter, and label data to ensure high-fidelity AI performance.

Labeling Experience

Education

Work History