Hugo Jácome Andrade

Key Skills

Software

Google Cloud Vertex AI

iMerit

Labelbox

Label Studio

Mercor

OneForma

Roboflow

Scale AI

Surge AI

Other

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Geospatial Tiled Imagery

Image

Text

Video

Top Label Types

Action Recognition

Audio Recording

Bounding Box

Classification

Computer Programming Coding

Data Collection

Emotion Recognition

Entity Ner Classification

Evaluation Rating

Fine Tuning

Function Calling

Geocoding

Mapping

Object Detection

Point Key Point

Polygon

Polyline

Prompt Response Writing SFT

Question Answering

Red Teaming

RLHF

Segmentation

Text Generation

Text Summarization

Tracking

Transcription

Translation Localization

Freelancer Overview

I am a specialist in RLHF for large language models and have extensive experience in image and video classification, data processing, and building high-quality datasets for AI systems. My background includes leading multidisciplinary teams to deliver AI-driven solutions across accessibility, computer vision, and multilingual technology projects. I have managed large-scale data annotation efforts, including localization into 48 languages and the creation of synthetic data for computer vision applications. My technical expertise spans from computer science, digital signal processing, and accessibility technologies, with a strong focus on developing inclusive, real-world AI products. I thrive in complex environments, breaking down challenges to first principles and driving innovation in data labeling, annotation, and training data pipelines.

Labeling Experience

Video specialized tasks

Don T DiscloseVideoSegmentationObject Detection

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of video-related tasks, some of them: - Object segmentation - Specialized captioning/transcription - OCR labeling - Video parameters and FPS analysis - Character recognition - Stick-to-reality parameters analysis

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of video-related tasks, some of them: - Object segmentation - Specialized captioning/transcription - OCR labeling - Video parameters and FPS analysis - Character recognition - Stick-to-reality parameters analysis

2021

Human speech labeling

Don T DiscloseAudioTranslation LocalizationEvaluation Rating

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of speech-related tasks, some of them: - Audio recording - Rubric creation and application for rating - Speech evaluation through dimensional rating - Specialized Transcription/Translation - Noise assessment - Factual checking - Safety assessment (red teaming) - Emotion recognition

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of speech-related tasks, some of them: - Audio recording - Rubric creation and application for rating - Speech evaluation through dimensional rating - Specialized Transcription/Translation - Noise assessment - Factual checking - Safety assessment (red teaming) - Emotion recognition

2020

RLHF for LLMs

Don T DiscloseTextEntity Ner ClassificationTranslation Localization

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of LLM-related tasks, some of them: - Rubric creation and application for rating - LLM evaluation and reinforcement learning - Specialized Translation/Localization - LLM classification tasks - Factual checking - Safety assessment (red teaming) - Emotion recognition

Different projects in labeling platforms working for AI frontier labs regarding AI RLHF, doing several types of LLM-related tasks, some of them: - Rubric creation and application for rating - LLM evaluation and reinforcement learning - Specialized Translation/Localization - LLM classification tasks - Factual checking - Safety assessment (red teaming) - Emotion recognition

2018

Sign Language Classification

Google Cloud Vertex AIVideoPoint Key PointSegmentation

The project was to launch an app able to provide sign language identification to close the communication gap between sign language users and those who don't natively understand it, so Deaf people can use their native language everywhere without the necessity of a 24/7 sign language interpreter when trying to communicate with everyone. Several types of labeling tasks were done, like segmentation, polygons, key-points, etc., in order to later feed the neural network. The app it's still using what I created and live in App Store: https://apps.apple.com/us/app/speakliz-for-deaf-people/id1239857536

The project was to launch an app able to provide sign language identification to close the communication gap between sign language users and those who don't natively understand it, so Deaf people can use their native language everywhere without the necessity of a 24/7 sign language interpreter when trying to communicate with everyone. Several types of labeling tasks were done, like segmentation, polygons, key-points, etc., in order to later feed the neural network. The app it's still using what I created and live in App Store: https://apps.apple.com/us/app/speakliz-for-deaf-people/id1239857536

2016 - 2023

Image classification & Object detection

Internal Proprietary ToolingImageBounding BoxSegmentation

The project was designed to launch a mobile app that performed advanced object detection and tracking with the goal of creating a useful digital daily assistant for Blind people by allowing them to have an entire audible description of their surroundings through the phone camera live feed. Therefore, with my team we created datasets for more than 500 objects (including from general ones, to even banknotes of +25 world currencies), OCR, distance measuring with phone camera, color detection, contextual places description, etc. Several types of labeling tasks were done, like bounding boxes, segmentation, polygons, etc., in order to later feed the neural network. The app it's still using what I created and live in App Store: https://apps.apple.com/us/app/vision-for-blind-people/id1444693229

The project was designed to launch a mobile app that performed advanced object detection and tracking with the goal of creating a useful digital daily assistant for Blind people by allowing them to have an entire audible description of their surroundings through the phone camera live feed. Therefore, with my team we created datasets for more than 500 objects (including from general ones, to even banknotes of +25 world currencies), OCR, distance measuring with phone camera, color detection, contextual places description, etc. Several types of labeling tasks were done, like bounding boxes, segmentation, polygons, etc., in order to later feed the neural network. The app it's still using what I created and live in App Store: https://apps.apple.com/us/app/vision-for-blind-people/id1444693229

2016 - 2023

Show more

Education

Work History