Hüseyin E.

Key Skills

Software

Scale AI

Top Subject Matter

AI Safety & Alignment

Advanced Logic & Technical Reasoning

Cognitive Behavioral Analysis

Top Data Types

Audio

Text

Image

Top Task Types

RLHF

Classification

Red Teaming

Fine-tuning

Transcription

Prompt + Response Writing (SFT)

Text Generation

Text Summarization

Evaluation/Rating

Question Answering

Function Calling

Entity (NER) Classification

Freelancer Overview

Bilingual (TR/EN) AI Evaluator and Alignment Specialist integrating cognitive psychology principles with advanced LLM training workflows. As a Quality Control Auditor at Scale AI, oversaw the entirety of the Turkish language data pipeline acting as the quality gatekeeper, while concurrently executing cross-domain evaluation as a CUA Tutor at xAI. Specialized in RLHF, SxS grading, adversarial Red Teaming, and multi-modal (Audio/Text) alignment. Demonstrated expertise in verifying function-calling logic and designing rule-based prompt architectures (e.g., intent branching) to strictly mitigate model hallucinations and sycophancy.

Labeling Experience

Coding User Action Tutor (xAI)

Computer Code ProgrammingFunction Calling

As a Coding User Action Tutor at xAI, I performed high-precision multi-modal evaluation and alignment auditing for software workflows, web development, audio, and rewriting tasks. I specialized in assessing complex function-calling, tool-use logic, spreadsheet productivity, and alignment of AI models within simulated environments. Additionally, I evaluated AI-generated code and technical reasoning, curated audio training data, and conducted style transfer rewriting to ensure consistent model behavior and high-fidelity outputs. • Conducted technical workflow optimization and evaluation in 'Macrohard' (LibreOffice Calc) environments. • Audited AI-generated web architecture and code for accuracy and instruction adherence. • Curated and aligned training data for advanced speech models and audio prompts. • Performed semantic refinement and style transfer to enhance naturalness and persona adherence.

2026 - Present

Quality Control Auditor (Scale AI)

Scale AITextRLHF

As a Quality Control Auditor & AI Trainer at Scale AI, I generated Golden Responses, conducted Side-by-Side (SxS) preference ranking, and executed RLHF-based evaluations in English and Turkish. My work included evaluating advanced model reasoning, function/API calls, RAG testing, and summarization capability over multi-turn chat scenarios. I also led adversarial red teaming, multi-modal data curation, and engineered custom rubrics for nuanced, cross-lingual evaluation in sensitive AI domains. • Generated preference rankings and performed bilingual SxS evaluation pipelines. • Evaluated long-context reasoning, tool-use, RAG, and summarization within chat environments. • Conducted safety red teaming against hallucination and sensitive content generation. • Curated high-fidelity visual and speech data and tailored evaluation rubrics for native alignment.

2025 - 2026

Education

Work History