Muneeb Ul haq

Key Skills

Software

Appen

CVAT

Data Annotation Tech

iMerit

Labelbox

Label Studio

Remotasks

Scale AI

SuperAnnotate

Surge AI

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Audio

Top Task Types

Polygon

Segmentation

Object Detection

Bounding Box

Tracking

Classification

Entity Ner Classification

Emotion Recognition

Translation Localization

Prompt Response Writing SFT

Question Answering

Text Generation

Text Summarization

Evaluation Rating

RLHF

Freelancer Overview

I am a detail-oriented professional with hands-on experience in AI data labeling, annotation, and large language model (LLM) content evaluation. My background includes evaluating and scoring AI-generated text and images using rubrics-based frameworks, working with Supervised Fine Tuning (SFT), RLHF, RLVR, and model alignment projects. I am skilled in text classification, sentiment analysis, entity tagging, and chatbot rating, with a strong focus on accuracy, consistency, and analytical reasoning. My multidisciplinary education in science and certifications in AI, data analysis, and cybersecurity enable me to adapt quickly to new tools and guidelines. I am passionate about contributing to high-quality AI training data and improving model performance in dynamic, remote work environments.

Labeling Experience

LLM Content Evaluator and AI Data Labeler

Scale AITextEvaluation Rating

I evaluated and rated LLM-generated content for quality, accuracy, and alignment using rubrics-based frameworks. Tasks included scoring outputs, analyzing content for safety and tone, and providing feedback for improvement in model responses. The experience involved Supervised Fine Tuning (SFT), RLHF, and RLVR workflows for comprehensive model assessment. • Evaluated chatbot and general AI outputs on criteria such as reasoning depth and prompt alignment. • Performed reinforcement learning-based reward signal evaluations to enhance alignment. • Labeled data across multiple batches, contributing directly to instruction dataset quality. • Utilized various online productivity and annotation tools for remote AI labeling work.

I evaluated and rated LLM-generated content for quality, accuracy, and alignment using rubrics-based frameworks. Tasks included scoring outputs, analyzing content for safety and tone, and providing feedback for improvement in model responses. The experience involved Supervised Fine Tuning (SFT), RLHF, and RLVR workflows for comprehensive model assessment. • Evaluated chatbot and general AI outputs on criteria such as reasoning depth and prompt alignment. • Performed reinforcement learning-based reward signal evaluations to enhance alignment. • Labeled data across multiple batches, contributing directly to instruction dataset quality. • Utilized various online productivity and annotation tools for remote AI labeling work.

2025

SparrowSignet, Instruction Writing & Rubric-Based Evaluation

Scale AITextText GenerationEvaluation Rating

Contributed to high-quality training data for instruction-tuned LLMs. Work included: Writing detailed, domain-diverse prompts and solutions. Creating structured rubrics for evaluating model responses. Performing rubric-based scoring for correctness, safety, and reasoning depth. Enhancing prompt clarity, diversity, and difficulty to meet evolving data standards.

Contributed to high-quality training data for instruction-tuned LLMs. Work included: Writing detailed, domain-diverse prompts and solutions. Creating structured rubrics for evaluating model responses. Performing rubric-based scoring for correctness, safety, and reasoning depth. Enhancing prompt clarity, diversity, and difficulty to meet evolving data standards.

2025

Aether, LLM Response Ranking & Evaluation

Scale AITextRLHFEvaluation Rating

Evaluated and ranked multiple AI-generated responses based on correctness, helpfulness, safety, and alignment with project guidelines. Responsibilities included: Scoring multi-turn conversations using fine-grained rubrics. Identifying safety risks, hallucinations, and policy violations. Selecting the best responses to train reward models (RLHF). Maintaining high consistency across edge cases and ambiguous queries.

Evaluated and ranked multiple AI-generated responses based on correctness, helpfulness, safety, and alignment with project guidelines. Responsibilities included: Scoring multi-turn conversations using fine-grained rubrics. Identifying safety risks, hallucinations, and policy violations. Selecting the best responses to train reward models (RLHF). Maintaining high consistency across edge cases and ambiguous queries.

2025

Guitar Riff, Creative Prompt Generation

Scale AITextText GenerationEvaluation Rating

Worked on generating diverse, high-quality creative prompts for a music-themed dataset used to train generative AI models. Tasks included: Writing original prompts for guitar riffs in multiple styles and moods. Calibrating difficulty and creativity levels according to project guidelines. Ensuring linguistic diversity, clarity, and model-compatible formatting. Reviewing and refining prompt–response pairs for coherence.

Worked on generating diverse, high-quality creative prompts for a music-themed dataset used to train generative AI models. Tasks included: Writing original prompts for guitar riffs in multiple styles and moods. Calibrating difficulty and creativity levels according to project guidelines. Ensuring linguistic diversity, clarity, and model-compatible formatting. Reviewing and refining prompt–response pairs for coherence.

2025

Guitar Pinstripe Prompt Generation, Text Polishing & Prompt Alignment (Outlier AI)

Surge AITextClassificationQuestion Answering

I am currently working on the Guitar Pinstripe project on Outlier AI, polishing and refining text responses so they sound natural and human. Edited grammar, fluency, tone, and creativity while ensuring every response strictly follows the assigned prompt. Evaluated whether the model fulfilled the instruction and corrected unnatural phrasing, punctuation, and off-prompt content. I also assisted in improving dataset quality for training large language models.

I am currently working on the Guitar Pinstripe project on Outlier AI, polishing and refining text responses so they sound natural and human. Edited grammar, fluency, tone, and creativity while ensuring every response strictly follows the assigned prompt. Evaluated whether the model fulfilled the instruction and corrected unnatural phrasing, punctuation, and off-prompt content. I also assisted in improving dataset quality for training large language models.

2025

Show more

Education

Work History