Leo M.

Key Skills

Software

Don't disclose

Top Subject Matter

LLM evaluation

safety/alignment and high-volume data annotation

Multimodal VQA evaluation for image-based AI outputs

Top Data Types

Text

Image

Document

Top Task Types

Question Answering

Tracking

Classification

Freelancer Overview

AI Evaluation & Data Annotation Specialist (Outlier / DataAnnotation / Alignerr). – Evaluated outputs from frontier language models — including Gemini, ChatGPT, and Grok — assessing reasoning quality, factual accuracy, and alignment with safety and policy guidelines. – Conducted structured data annotation across a wide range of technical and analytical task categories, consistently meeting quality benchmarks under high-volume conditions. – Identified and documented prompt inconsistencies, logical errors, and model performance gaps; produced detailed, structured feedback that fed directly into model improvement pipelines. – Operated under strict project confidentiality requirements, demonstrating disciplined adherence to annotation guidelines and quality protocols throughout all engagements. – Evaluated visual question-answering (VQA) tasks and model responses at Alignerr, verifying accuracy and correctness of AI outputs against image-based queries; identified errors and supplied corrected responses to support multimodal model improvement. Graduating with a bachelor from Social Work & Social Sciences – Core modules: Sociology, Social Economics, Descriptive Social Statistics, Political Sociology, and Human Rights. – Fieldwork: Conducted structured client intake interviews, needs assessments, and case documentation within local social institutions.

Labeling Experience

AI Evaluation & Data Annotation Specialist

ImageTrackingClassification

No description provided.

2025 - Present

AI Evaluation & Data Annotation Specialist (Alignerr) - VQA verification

Don't discloseImageQuestion Answering

Evaluated visual question-answering (VQA) tasks and model responses against image-based queries. Checked correctness and accuracy of multimodal outputs and reported issues found during review. • Assessed answer validity relative to image content • Logged errors and inconsistencies for follow-up • Supplied corrected responses when needed • Contributed to multimodal model improvement efforts

2025 - Present

AI Evaluation & Data Annotation Specialist (Outlier / DataAnnotation / Alignerr)

Don't discloseText

Assessed outputs from frontier language models (Gemini, ChatGPT, and Grok) for quality and compliance. Focus areas included reasoning quality, factual accuracy, and alignment with safety and policy guidelines. • Measured response quality against established benchmarks • Identified prompt inconsistencies and logical errors • Documented model performance gaps for improvement • Supplied structured feedback for model improvement pipelines

2025 - Present

Education

Work History