Description

AI Data Specialist with 5+ years of experience improving Speech AI and Search quality through high-precision human evaluation (RLHF). I curate and refine multilingual (French/English) datasets, run side-by-side model comparisons, and deliver publication-ready transcripts (verbatim or clean-read) for R&D and media use. I’m rigorous about text normalization (numbers, units, abbreviations), dataset consistency at scale, and complex audio scenarios (multi-speaker, overlaps, hesitations, ambient noise).

On top of data work, I also build fast MVPs via AI-assisted prototyping (“vibe coding”) for web & mobile apps—turning ideas into working prototypes quickly, with clear iterations and practical deliverables.

Finally, I bring a strong creative edge: I’m a published poetry author and I create content around philosophy and ideas. That means I care deeply about language, tone, and clarity—useful for UX copy, localization, and overall output quality.

If you need reliable speech/ASR data curation, transcription QA, search relevance labeling, or rapid app prototyping, I can help.

Industry field of expertise

Languages

French
Native or bilingual
English
Fluent

Workplace preferences

Remote only

Primarily works remotely

ElevenLabs
AI Speech Data Specialist
CIVIC AND SOCIAL ORGANIZATIONS
January 2021 - Today (5 years and 5 months)
Curated and refined training data for AI speech recognition models, contributing directly to the improvement of AI generated transcription systems.
• • Evaluated and corrected AI generated transcriptions for ElevenLabs, improving output accuracy for speech model training pipelines
• • Processed dozens of audio hours monthly, maintaining consistent quality under strict production deadlines
• •
Applied systematic text normalization protocols (numbers, abbreviations, units) ensuring dataset consistency across thousands of entries
• • Managed complex audio scenarios: multi-speaker identification, overlapping speech, hesitations, and ambient noise annotation
• • Produced publication-ready verbatim and clean-read transcripts for speech recognition R&D and media applications
intelligence artificielle Transcription
Appen
AI Search Quality Evaluator & Data Annotator
CIVIC AND SOCIAL ORGANIZATIONS
January 2021 - January 2023 (2 years)
Provided human evaluation data used to train and improve search ranking algorithms, directly contributing to AI model performance.
• •
Analyzed and labeled search queries and results based on relevance, user intent, and source quality
• • Identified and flagged data anomalies — outdated, duplicate, misleading, or low-quality content
• • Documented edge cases and ambiguous patterns to support refinement of evaluation rubrics (directly analogous to AI red-teaming)
• • Consistently met or exceeded daily quality and productivity quotas on proprietary annotation platforms