Maria Catalina B.

Key Skills

Software

Clickworker

Data Annotation Tech

Figure Eight

Labelbox

Lionbridge

OneForma

Scale AI

Sloth

Telus

Appen

Mercor

Top Subject Matter

AI safety training and evaluation (LLM alignment)

AI adversarial testing and safety evaluation

LLM red teaming and safety evaluation

Top Data Types

Audio

Document

Medical Dicom

Text

Top Task Types

Action Recognition

Classification

Evaluation/Rating

Red Teaming

Translation/Localization

RLHF

Transcription

Freelancer Overview

Team Leader – RLHF & SFT Projects (Mercor | Remote). Brings 16+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Mercor, N, and A. Education includes Bachelor’s Degree, University of Letters “Alexandru Ioan Cuza”. AI-training focus includes data types such as Text, Medical, and DICOM and labeling workflows including RLHF, Red Teaming, and Evaluation.

Labeling Experience

Team Leader – RLHF & SFT Projects (Mercor | Remote)

MercorTextRLHF

Supervised annotators and reviewers to produce high-quality training data for RLHF and SFT. Monitored workflow quality, productivity, and project compliance across multiple AI training initiatives. Performed quality assurance reviews and resolved complex edge cases in instruction-following and AI outputs. • Supervised human feedback collection workflows • Updated annotation guidelines with stakeholders • Conducted QA on edge cases and evaluation consistency • Mentored team members and improved throughput

2026 - Present

Red Teaming Team Leader (Innodata | Remote)

TextRed Teaming

Coordinated a team of AI red teamers for specialized adversarial testing projects. Developed and assigned adversarial testing strategies to uncover model vulnerabilities and safety weaknesses. Consolidated results into detailed vulnerability and risk assessment reports for stakeholders. • Mentored on jailbreaking techniques and ethical testing practices • Designed prompt engineering approaches for robustness checks • Ensured timelines and quality standards • Produced detailed findings for decision-makers

2025 - Present

Medical Transcriptionist (Remote)

Transcription

Transcribed and edited complex medical reports, patient histories, and discharge summaries with high accuracy. Applied medical terminology and pharmacology knowledge while maintaining confidentiality and compliance. Collaborated with healthcare professionals to clarify inconsistencies and improve record quality. • Produced accurate typed medical documentation • Ensured consistent terminology usage • Maintained privacy and compliance standards • Resolved content discrepancies with clinicians

2024 - Present

AI Red Teaming Specialist (Remote)

TextRed Teaming

Conducted adversarial testing on large language models to identify vulnerabilities, biases, and safety risks. Designed and executed jailbreak prompts to evaluate robustness of AI safety guardrails. Analyzed model behavior and reported harmful, misleading, or inaccurate outputs. • Performed jailbreak and robustness evaluations • Produced detailed analysis reports • Contributed to alignment and safety improvement • Assessed harmfulness and bias patterns

2023 - Present

Audio data labelling

AppenAudioClassificationQuestion Answering

Labeled and transcribed audios for YouTube and gemini voice projects, helping improve voice recognition and conversational AI. Tagged speaker intent, tone and context.

2016

Show more

Education

Work History