Open to opportunities

Sofia Melissa Cerna Meneses

@sofiamelissacernamen

Message

I evaluate and train multimodal AI models through RLHF, SFT, and quality auditing.

Peru

Message

What I'm looking for

I'm looking for remote AI data, multimodal evaluation, RLHF, SFT, and model-alignment work where I can deliver rigorous, localized, high-quality datasets across text, audio, image, and video.

I've delivered AI evaluation and annotation work for Mercor, RWS Group, Lilt, Trustscale, Invisible Technologies, and Telus Digital, training and evaluating LLMs and vision-language models across text, audio, image, and video. My work includes RLHF evaluation, SFT dataset creation, prompt engineering, pairwise comparison, hallucination auditing, and rubric-based scoring.

I've completed 1,000+ multi-constraint prompt evaluations with 98%+ quality adherence, including 500+ end-to-end multimodal evaluations and 300+ speech-to-speech and multimodal outputs. I also bring digital media analytics and dashboard experience from Publicis Global Delivery, where I support campaign setup, pacing, bid optimization, and performance reviews.

Experience

Work history, roles, and key accomplishments

Current

AI Data Generalist

Current

Mercor

Jul 2026 - Present (2 months)

Executed 500+ end-to-end multimodal model evaluations on macOS, testing instruction-following, tone adaptation, and formatting. Conducted pairwise comparisons and authored evidence-based rationales for RLHF datasets.