I’m Marloes Gerdes, a native Dutch annotation and evaluation specialist working on training data and model outputs for large-scale AI products. I run pairwise model comparisons, rate outputs on defined quality scales, and categorize labeling errors against documented taxonomies—producing gold-standard Dutch reference responses and keeping the same judging rigor consistent across large volumes of items. In addition to evaluation, I build and deploy production LLM applications and support them as someone who understands how outputs are generated, not just how they read. My focus includes bilingual Dutch/English delivery (CEFR C2), daily quality review of contributor work, and practical implementation with GDPR-minded setups for healthcare and e-commerce contexts, including building assistant workflows, triage support, and public demo suites.

Marloes Gerdes

I’m Marloes Gerdes, a native Dutch annotation and evaluation specialist working on training data and model outputs for large-scale AI products. I run pairwise model comparisons, rate outputs on defined quality scales, and categorize labeling errors against documented taxonomies—producing gold-standard Dutch reference responses and keeping the same judging rigor consistent across large volumes of items. In addition to evaluation, I build and deploy production LLM applications and support them as someone who understands how outputs are generated, not just how they read. My focus includes bilingual Dutch/English delivery (CEFR C2), daily quality review of contributor work, and practical implementation with GDPR-minded setups for healthcare and e-commerce contexts, including building assistant workflows, triage support, and public demo suites.

Available to hire

I’m Marloes Gerdes, a native Dutch annotation and evaluation specialist working on training data and model outputs for large-scale AI products. I run pairwise model comparisons, rate outputs on defined quality scales, and categorize labeling errors against documented taxonomies—producing gold-standard Dutch reference responses and keeping the same judging rigor consistent across large volumes of items.

In addition to evaluation, I build and deploy production LLM applications and support them as someone who understands how outputs are generated, not just how they read. My focus includes bilingual Dutch/English delivery (CEFR C2), daily quality review of contributor work, and practical implementation with GDPR-minded setups for healthcare and e-commerce contexts, including building assistant workflows, triage support, and public demo suites.

See more

Language

English
Fluent
Dutch
Fluent

Work Experience

AI Annotation, Evaluation & Review Specialist — Dutch (English to Dutch) at TeachX AI Programme
March 1, 2026 - Present
Daily quality review of other contributors’ Dutch outputs against project guidelines, holding official review responsibility alongside my own annotation work. Perform pairwise model evaluations by comparing responses side-by-side, selecting the stronger output, and documenting preference reasoning for model training and validation. Rate model outputs using defined quality scales and category error taxonomies, authoring gold-standard Dutch reference responses. Also includes audio/speech labeling assessment (transcription accuracy, speech recognition errors, and correcting generated responses for naturalness/locale/register), and image/document labeling. Maintain Dutch conventions across deliverables (register, terminology, product naming, and formatting) and ensure consistency with the reference standard across product tiers.
AI Implementation Specialist (independent) at Captured by AI
December 1, 2024 - Present
Design and build production LLM applications for Dutch healthcare, cosmetics, dental, and physiotherapy practices—from process analysis and discovery through deployment and ongoing support. Regularly evaluate and correct model outputs in Dutch for systems built by me, applying the same judgment criteria as contract evaluation work. Work directly with the Claude API and Anthropic vision models and Claude Code, deploying on Cloudflare Workers with Supabase, with GDPR/AVG requirements treated as a starting condition (secure staff dashboards, controlled handling of patient data, and automated logging). Build assistant products including a live Dutch ENT clinic triage assistant backed by Supabase on Cloudflare Workers, addressing security gaps (authentication, self-registration, and BSN validation) and migrating from no-code prototyping to a maintainable Claude Code workflow.
SEO & E-commerce Specialist at L'Atelier
October 1, 2022 - October 1, 2025
Owned on-page SEO and content strategy for the web shop and managed product content across a catalog of 100+ product pages.
Digital Marketeer at Meertens Meubelen
January 1, 2015 - October 1, 2022
On-Page SEO optimisation; Content marketing & strategy; Product content management; Customer service.

Education

Bachelor of Social Work at Hanzehogeschool Groningen
January 1, 1997 - January 1, 2001

Qualifications

Google: AI Essentials
February 1, 2026 - March 6, 2026
IBM: GenAI for SEO: a hands-on playbook
February 1, 2026 - March 6, 2026
IBM: Search Engine Optimization and Content Marketing
February 1, 2026 - March 6, 2026
Google: AI Performance Based Ads
January 1, 2026 - March 6, 2026
Bit Academy: ChatGPT
December 1, 2024 - March 6, 2026
Bit Academy: Prompt Engineering
October 1, 2024 - March 6, 2026
Bit Academy: HTML/CSS Beginner
September 1, 2024 - March 6, 2026
Winc Academy: Search Engine Optimization
January 1, 2023 - March 6, 2026
Anthropic — Claude Code 101
January 1, 2026 - January 1, 2026
Anthropic — Claude 101
January 1, 2026 - January 1, 2026
Anthropic — Introduction to Agent Skills
January 1, 2026 - January 1, 2026
Anthropic — Introduction to Claude Cowork
January 1, 2026 - January 1, 2026
Google — AI Essentials
January 1, 2026 - January 1, 2026
IBM — GenAI for SEO
January 1, 2026 - January 1, 2026
LinkedIn — AI Foundations: Machine Learning
January 1, 2026 - January 1, 2026
Attendee, Code with Claude (London)
May 1, 2026 - May 1, 2026
Claude Code architecture certification (in progress)
January 11, 2030 - September 8, 2026

Industry Experience

Software & Internet, Professional Services, Media & Entertainment, Retail, Education, Healthcare