Arman Hossain

Key Skills

Software

Appen

Data Annotation Tech

Telus

Labelbox

Top Subject Matter

Nlp Domain Expertise

Llms Domain Expertise

General AI

Top Data Types

Document

Image

Video

Text

Top Task Types

Emotion Recognition

Evaluation/Rating

Mapping

Object Detection

Segmentation

Classification

Freelancer Overview

AI Trainer and QA - Telus, Appen, RWS, Outlier AI. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Appen and Labelbox. Education includes Certificate, Harvard University (edX) and Higher Secondary Diploma, Chhotajagulia High School. AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and Classification.

Labeling Experience

Instagram image entity tagging

TelusImageSegmentationClassification

Analyzed Instagram images to identify and tag all visible entities, including people, clothing, style, animals, locations, and more. Each entity was paired with alternate reference images to provide rich context. This structured labeling helps AI models understand visual content, recognize objects and people accurately, and improve image-based recommendations, content moderation, and visual search applications.

Analyzed Instagram images to identify and tag all visible entities, including people, clothing, style, animals, locations, and more. Each entity was paired with alternate reference images to provide rich context. This structured labeling helps AI models understand visual content, recognize objects and people accurately, and improve image-based recommendations, content moderation, and visual search applications.

2025 - 2025

AI Video Prompt Reverse Engineering

TelusVideoClassificationEmotion Recognition

Examined Instagram videos to deconstruct AI-generated or inspired content, identifying key visual and temporal elements, scene compositions, actions, and styles to infer underlying prompts. This enabled AI to generate or recreate videos with greater accuracy and fidelity. Supports AI in understanding temporal dynamics and multi-frame content generation, enhancing applications in video synthesis, creative media, motion tracking, real-time animation, and content personalization.

Examined Instagram videos to deconstruct AI-generated or inspired content, identifying key visual and temporal elements, scene compositions, actions, and styles to infer underlying prompts. This enabled AI to generate or recreate videos with greater accuracy and fidelity. Supports AI in understanding temporal dynamics and multi-frame content generation, enhancing applications in video synthesis, creative media, motion tracking, real-time animation, and content personalization.

2025

AI Image Prompt Reverse Engineering

TelusTextClassificationText Generation

Analyzed Instagram images to reverse-engineer prompts used in AI-generated content, identifying visual elements, styles, compositions, and attributes that influenced generation. This enabled precise reconstruction of prompts to guide AI for consistent and high-quality outputs. Helps AI models learn correlations between textual prompts and visual outputs, improving image generation, content creation tools, design assistance, and visual AI research.

Analyzed Instagram images to reverse-engineer prompts used in AI-generated content, identifying visual elements, styles, compositions, and attributes that influenced generation. This enabled precise reconstruction of prompts to guide AI for consistent and high-quality outputs. Helps AI models learn correlations between textual prompts and visual outputs, improving image generation, content creation tools, design assistance, and visual AI research.

2025

Processed Instagram videos to detect and tag all entities frame by frame, including people, attire, style, animals, locations, and actions. Entities were cross-referenced with alternate images to enhance AI understanding. This enables AI to interpret dynamic content, track objects and people over time, and improve applications in video summarization, real-time recognition, and content analysis.

TelusVideoClassificationObject Detection

Processed Instagram videos to detect and tag all entities frame by frame, including people, attire, style, animals, locations, and actions. Entities were cross-referenced with alternate images to enhance AI understanding. This enables AI to interpret dynamic content, track objects and people over time, and improve applications in video summarization, real-time recognition, and content analysis.

Processed Instagram videos to detect and tag all entities frame by frame, including people, attire, style, animals, locations, and actions. Entities were cross-referenced with alternate images to enhance AI understanding. This enables AI to interpret dynamic content, track objects and people over time, and improve applications in video summarization, real-time recognition, and content analysis.

2025

AI Data Annotation & LLM Data Training (Personal Projects and Freelance)

LabelboxTextClassification

Conducted personal data annotation and LLM data training over a period exceeding three years. Responsibilities included tagging, cleaning, and preparing text data for AI model development and evaluation. Maintained consistent data quality and accuracy across multiple annotation software tools. • Specialized in entity tagging, sentiment analysis, and emotion recognition tasks. • Used platforms such as Labelbox, SuperAnnotate, Appen, and Telus International. • Focused on prompt engineering, QA validation, and hands-on LLM model fine-tuning. • Ensured 95–100% data accuracy through methodical quality assurance processes.

Conducted personal data annotation and LLM data training over a period exceeding three years. Responsibilities included tagging, cleaning, and preparing text data for AI model development and evaluation. Maintained consistent data quality and accuracy across multiple annotation software tools. • Specialized in entity tagging, sentiment analysis, and emotion recognition tasks. • Used platforms such as Labelbox, SuperAnnotate, Appen, and Telus International. • Focused on prompt engineering, QA validation, and hands-on LLM model fine-tuning. • Ensured 95–100% data accuracy through methodical quality assurance processes.

2021 - Present

Show more

Education

Work History