Muhammad Saqib

Key Skills

Software

Appen

Clickworker

CrowdSource

Humanatic

Toloka

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Video

Top Task Types

Audio Recording

Computer Programming Coding

Evaluation Rating

Object Detection

Prompt Response Writing SFT

Freelancer Overview

I am a hands-on AI Training specialist with a deep background in high-stakes computer vision and multilingual LLM evaluation. most of my work has focused on providing the "ground truth" for Autonomous systems, where I have spent hundreds of hours labeling complex self-driving car imaging. I don't just draw boxes; I focus on the difficult edge cases like low-light pedestrians or occluded obstacles because I know that real-world model safety depends on the human precision I provide. One of my most significant contributions was as a core specialist for Project Fireweed, where I focused on LLM localization and multi-language evaluation. I acted as a subject matter expert to train model engines on identifying linguistic nuances and generating contextually fluent responses. Additionally, I have extensive experience in industrial image classification, having trained engines to identify material attributes like wood species and quality grades from unstructured data.

Labeling Experience

Multilingual LLM Localization

AppenTextText GenerationRLHF

Contributed to Project Fireweed, an Appen initiative aimed at enhancing large language models (LLMs) for better performance in diverse languages and regional dialects. Specialized in generating high-quality prompt-response pairs, multi-turn conversations, and localized questions to train AI chatbots for factual, logical, contextually relevant, and culturally appropriate outputs. Evaluated and rated AI-generated responses for helpfulness, fluency, safety, and hyper-local relevance, while adhering to strict guidelines on multi-turn interactions and difficulty levels (easy/medium/hard). Processed timed tasks efficiently in the ADAP platform, focusing on improving chatbot communication in underrepresented languages/dialects. Maintained high-quality standards to support safer and more inclusive generative AI models.

Contributed to Project Fireweed, an Appen initiative aimed at enhancing large language models (LLMs) for better performance in diverse languages and regional dialects. Specialized in generating high-quality prompt-response pairs, multi-turn conversations, and localized questions to train AI chatbots for factual, logical, contextually relevant, and culturally appropriate outputs. Evaluated and rated AI-generated responses for helpfulness, fluency, safety, and hyper-local relevance, while adhering to strict guidelines on multi-turn interactions and difficulty levels (easy/medium/hard). Processed timed tasks efficiently in the ADAP platform, focusing on improving chatbot communication in underrepresented languages/dialects. Maintained high-quality standards to support safer and more inclusive generative AI models.

2024 - 2025

Education

Work History