João Vitor L.

Key Skills

Software

Labelbox

Lionbridge

OneForma

Scale AI

Telus

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Video

Top Task Types

Action Recognition

Audio Recording

Bounding Box

Object Detection

RLHF

Freelancer Overview

I am an AI training and evaluation specialist with hands-on experience improving generative AI systems through prompt engineering, LLM evaluation, and high-quality data annotation. I have worked on multiple AI training projects involving text, code, and multimodal data, following strict quality guidelines to ensure consistency, accuracy, and usefulness of model outputs. My background includes reviewing and debugging code (Python, C++, Java), creating and refining prompts for web-based reasoning tasks, and labeling text, image, video, and audio data in English using industry tools such as Labelbox. I bring a strong analytical mindset, attention to detail, and a solid technical foundation, allowing me to contribute effectively to human-in-the-loop workflows and AI model improvement at scale.

Labeling Experience

AI Task Designer & Prompt Engineering Specialist

Internal/Proprietary ToolingComputer Code ProgrammingText GenerationRLHF

Worked on multiple AI training projects as both an attempter and reviewer, contributing to the full human-in-the-loop evaluation cycle for large language models. As an attempter, designed high-quality prompts aligned with predefined metadata (domain, intent, difficulty, and language) and produced structured, step-by-step attempts for coding and web navigation tasks, ensuring realistic and testable scenarios. As a reviewer, developed detailed evaluation rubrics with objective criteria covering correctness, code quality, scalability, completeness, and clarity. Evaluated AI-generated responses and human attempts by inspecting reasoning steps, executing code in virtual machine environments when necessary, and validating web navigation behavior in browser-based environments. Ranked responses from best to worst with clear technical justification, identifying edge cases, failure patterns, and reasoning gaps to support continuous model improvement.

2025

AI Training Evaluator & Code Review Specialist

Scale AIComputer Code ProgrammingEvaluation/RatingComputer Programming/Coding

Worked as an AI training evaluator on multiple generative AI projects, contributing to the evaluation, refinement, and improvement of large language model outputs. Responsibilities included prompt creation and optimization, response evaluation and ranking, and detailed analysis of model behavior across diverse tasks involving natural language, code, and reasoning. Additionally, performed code review and debugging tasks in Python, C++, and Java to validate correctness, efficiency, and adherence to best practices within AI-generated solutions. Contributed to data annotation and quality assurance processes, following strict guidelines to ensure consistency, accuracy, and usefulness of training data within human-in-the-loop workflows.

2024

Multimodal Data Annotation Specialist (Video, Audio, Image)

LabelboxVideoBounding BoxAction Recognition

Worked as a data annotator and reviewer on multiple short- and mid-term multimodal AI training projects. Responsibilities included detailed video event annotation for sports content, such as identifying and validating key football match events (e.g., ball crossing the goal line, goal kicks, goalkeeper possession, referee card events, and assistant referee flag signals) using temporal segmentation and frame-level accuracy. Also contributed to media annotation projects involving films and TV series, where tasks included scene change detection, character transition labeling, dialogue segmentation, background music identification, and audio-to-text alignment. Additional projects involved diverse multimodal annotation tasks across video, audio, image, and text data, following strict quality guidelines and consistency checks to support human-in-the-loop AI model training workflows.

2025 - 2025

Education

Work History