Reda Elboukfaoui

Key Skills

Software

Appen

Clickworker

Mercor

OneForma

Remotasks

Scale AI

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Task Types

Classification

Evaluation Rating

Text Generation

Text Summarization

Freelancer Overview

I have started data labeling and annotation tasks 2 years ago. I've contributed to text and audio LLM training and RLHF projects on Outlier.ai like Cypher evals c, Pangolin Safety or Guitar Riff. The fact that I am bilingual helped me a lot through French and English projects. Besides, I have a background in Information Systems Engineering.

Labeling Experience

Pangoling Safety

Scale AITextClassificationFine Tuning

The project aims to refine the refusal and safety mechanisms of AI models. It ensures that when a user asks for something dangerous (e.g., illegal instructions, hate speech, or self-harm), the model responds safely, neutrally, and according to strict ethical guidelines without being overly restrictive on benign topics. Sub-projects: - Pangolin Text: (mostly) Focusing on written dialogue and reasoning. - Pangolin Vision: Evaluating how AI interprets and responds to sensitive or unsafe images.

The project aims to refine the refusal and safety mechanisms of AI models. It ensures that when a user asks for something dangerous (e.g., illegal instructions, hate speech, or self-harm), the model responds safely, neutrally, and according to strict ethical guidelines without being overly restrictive on benign topics. Sub-projects: - Pangolin Text: (mostly) Focusing on written dialogue and reasoning. - Pangolin Vision: Evaluating how AI interprets and responds to sensitive or unsafe images.

2025 - 2025

Cypher evals C

Scale AITextClassificationQuestion Answering

Contributors (typically software developers or computer science students) generally perform the following: - Evaluation (RLHF): Comparing two AI-generated responses to a C programming prompt and determining which is more accurate, secure, and efficient. - Writing "Gold Standards": Authoring perfect C solutions to serve as reference data for the model. - Debugging: Identifying and fixing complex errors (memory leaks, buffer overflows, segmentation faults) in code produced by the AI.

Contributors (typically software developers or computer science students) generally perform the following: - Evaluation (RLHF): Comparing two AI-generated responses to a C programming prompt and determining which is more accurate, secure, and efficient. - Writing "Gold Standards": Authoring perfect C solutions to serve as reference data for the model. - Debugging: Identifying and fixing complex errors (memory leaks, buffer overflows, segmentation faults) in code produced by the AI.

2024 - 2025

Education

Work History