Open to opportunities

Ayuk Mostel

@ayukmostel

Message

AI Data Annotation Specialist delivering zero-error, high-fidelity labels for LLM alignment and integrity.

Cameroon

Message

What I'm looking for

I’m looking for high-stakes AI projects where I can audit model reasoning, label image/video/text with forensic QA, and harden training data using RLHF and reproducible error traces to drive measurable performance gains.

I’m an AI Data Annotation Specialist with a foundation in Systems Architecture and Technical Forensics. I bridge the gap between raw data and machine intelligence using high-precision annotation and labeling that supports Large Language Models (LLMs).

I’m known for identifying false signals in datasets and providing the high-quality human feedback (RLHF) needed for model alignment. With a forensic approach, I harden semantics, correct logical hallucinations, and apply bias mitigation to improve training honesty and consistency.

Across international, high-volume work, I’ve maintained 100% diagnostic accuracy and zero-error drift while managing high-density workloads (200+ units). I also use Reproducible Error Trace (RET) methodology to document and resolve systemic failures in automated data.

Previously, I led scalable technical systems deployment—standardizing configurations and deploying 200+ workstations—then translating that same diagnostic rigor into AI data quality. I’ve performed image and video annotation on Toloka and Atlas, and I continue model integrity testing through Generative AI auditing and LLM evaluation.

Experience

Work history, roles, and key accomplishments

Current

Bilingual Logic Auditor

Current

Global Markets

Jan 2021 - Present (5 years 5 months)

Provided rigorous accuracy checks on AI reasoning traces for international projects, ensuring high-signal training data output. Used the Reproducible Error Trace (RET) method to document and resolve systemic failures for cleaner datasets.

Data Labeling Forensic Documentation (RET)Bilingual Technical SME

Current

Technical Systems Deployment Lead

Current

Isiss SARL

Jan 2021 - Present (5 years 5 months)

Orchestrated standardized deployment of 200+ workstations for high-compliance institutions with a zero-failure mindset. Performed proactive hardware/software hardening that reduced system downtime by 40% and completed image/video annotations on Toloka and Atlas for AI model training pipelines.

Systems Deployment Quality Control Root Cause Analysis Troubleshooting Video Annotation Atlas

Education

Degrees, certifications, and relevant coursework

LinkedIn Learning

LinkedIn Learning Certificate, Generative AI

2026 -

Completed a LinkedIn Learning course on Generative AI concepts. Applied generative AI evaluation skills to identify where models can hallucinate in high-stakes contexts.