Daniel Mulwa

Key Skills

Software

Other

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Video

Text

Top Label Types

Bounding Box

Text Generation

Evaluation Rating

Prompt Response Writing SFT

Text Summarization

Freelancer Overview

I am an experienced data labeling and evaluation specialist with a strong background in annotating and assessing AI training data across text, image, audio, and video domains. My work on projects like Handshake (Project Hedgehog) and Prolific has involved labeling data for supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF), evaluating AI-generated content for accuracy, bias, and safety, and simulating real-world tasks to test model performance. I am skilled in providing detailed human judgment, fact-checking, and identifying model inaccuracies, with a keen eye for quality and logical coherence. My technical expertise is complemented by a solid foundation in data analysis, visualization, and research, using tools such as Python, R, and SQL, and I am adept at communicating findings to both technical and non-technical audiences.

Labeling Experience

Annotator

OtherVideoBounding BoxText Generation

The project tasks involved annotating video outputs to train the LLM on video generation with prompts and evaluating or comparing AI-generated responses by identifying inaccuracies and bias in model outputs. Project Hedgehog also entailed labeling data for SFT (Supervised Fine-Tuning) and RLHF (Reinforcement Learning from Human Feedback) and providing human judgment to help models evolve in efficiency.

The project tasks involved annotating video outputs to train the LLM on video generation with prompts and evaluating or comparing AI-generated responses by identifying inaccuracies and bias in model outputs. Project Hedgehog also entailed labeling data for SFT (Supervised Fine-Tuning) and RLHF (Reinforcement Learning from Human Feedback) and providing human judgment to help models evolve in efficiency.

2025 - 2025

Data Labelling and Evaluation Participant

Data Annotation TechTextBounding BoxText Generation

The projects involved data annotation and labeling across text, image, audio, and video for training supervised models, Reinforcement Learning from Human Feedback (RLHF) where I used to score, rank, and compare AI model outputs based on quality, tone, or accuracy. There was also content evaluation and fact-checking that included assessing AI-generated content for safety, bias, logical coherence, and factual accuracy against trusted sources. Lastly, interaction and task simulation where I could engage the LLM in a real-world conversational tasks like email drafting, querying, prompting and and problem-solving to train and test its performance.

The projects involved data annotation and labeling across text, image, audio, and video for training supervised models, Reinforcement Learning from Human Feedback (RLHF) where I used to score, rank, and compare AI model outputs based on quality, tone, or accuracy. There was also content evaluation and fact-checking that included assessing AI-generated content for safety, bias, logical coherence, and factual accuracy against trusted sources. Lastly, interaction and task simulation where I could engage the LLM in a real-world conversational tasks like email drafting, querying, prompting and and problem-solving to train and test its performance.

2023 - 2025

Education

Work History