Mario H.

Key Skills

Software

Labelbox

Scale AI

Don't disclose

Top Subject Matter

LLM training and software engineering model evaluation

Top Data Types

Computer Code Programming

Text

Top Task Types

Computer Programming/Coding

Data Collection

Prompt + Response Writing (SFT)

Text Summarization

Freelancer Overview

AI Engineer / LLM Training Specialist (Expert data labeling, prompt/response & code evaluation, task artifact developmen. Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose. Education includes Bachelor of Science in Engineering, Catholic University of Maule (2023). AI-training focus includes data types such as Computer Code and Programming and labeling workflows including Evaluation and Rating.

Labeling Experience

AI Engineer / LLM Training Specialist (Expert data labeling, prompt/response & code evaluation, task artifact development)

Don't disclose

Provided expert data labeling and evaluation for LLM training focused on software engineering, code generation, debugging, reasoning, tool use, and agentic workflows. Assessed model-generated prompts, responses, code diffs, unit tests, logs, and execution results using strict technical rubrics. • Evaluated frontier LLM behavior across coding and reasoning scenarios • Identified failure modes including hallucinations, missing edge cases, and architectural weaknesses • Performed senior-level checks on correctness, maintainability, security, performance, and debugging strategy • Developed task artifacts such as programming tasks and validation materials

2025 - Present

Outlier AI

Scale AIComputer Code ProgrammingRLHFEvaluation/Rating

I worked on a large-scale AI training data project involving the evaluation and refinement of LLM-generated outputs for code generation, multi-turn prompts, and reasoning tasks. The scope included creating high-quality training data through rigorous rubric-based review, validation of test logs, and refinement of prompts for intent clarity and instructional alignment. Tasks included reviewing code diffs, validating unit test coverage based on execution logs, rewriting problem statements and requirements to match original inputs, and scoring outputs using structured evaluation criteria. The project operated at scale, with hundreds of daily annotations and peer-reviewed submissions. Quality standards were strict, including log-based test validation, rubric adherence, and multi-stage review workflows to ensure consistency, fairness, and reproducibility.

2024

Education

Work History