Michael Ramirez - Specialist in Data Labeling and Annotation for AI Training and Developmen

Key Skills

Software

LabelImg

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Geospatial Tiled Imagery

Text

Top Task Types

Bounding Box

Data Collection

Question Answering

Freelancer Overview

I am an experienced professional in Artificial Intelligence, with a strong background in developing innovative solutions for geospatial and technical applications. I have designed and implemented a geographic viewer that predicts TEC (Total Electron Content) values based on historical data using a predictive model, providing insights into atmospheric conditions through advanced machine learning algorithms. Additionally, I developed a neural network capable of processing satellite images by segmenting them into partitions and training the model to identify and compare cities. This project demonstrated my expertise in image recognition and geospatial analysis, leveraging AI to derive meaningful insights from visual data. I also created a geodesy-focused chatbot using a question-answer model to provide expert-level responses to technical and theoretical queries in the field. This solution integrates natural language processing (NLP) to deliver accurate and efficient knowledge retrieval for users. My work combines advanced AI techniques, including predictive modeling, image recognition, and NLP, to solve complex problems and optimize workflows, making me a valuable asset in AI and geospatial innovation projects.

IntermediateEnglishSpanish

Labeling Experience

TEC Prediction Geographic Viewer

Internal Proprietary ToolingGeospatial Tiled ImageryComputer Programming Coding

Developed a geographic viewer to predict Total Electron Content (TEC) values by annotating and processing historical geospatial data. This project serves as my thesis project for my degree in Cadastral and Geodetic Engineering. Designed and implemented a predictive AI model using TensorFlow to analyze atmospheric variations. This project required precise labeling of data to train the model effectively, optimizing predictions for decision-making processes. Tools like Python and JavaScript were used to integrate interactive visualization, ensuring accuracy and usability for end-users.

2024

Geodesy Question-Answer Chatbot

Internal Proprietary ToolingTextQuestion Answering

Created a geodesy-focused chatbot using a question-answer model to provide expert responses to technical and theoretical queries. Curated and labeled a dataset of geodesy-related questions and answers to train the model effectively. Implemented the chatbot using natural language processing (NLP) with TensorFlow, enabling precise and efficient support for geospatial professionals.

2023 - 2023

Satellite Image Neural Network for City Identification

Internal Proprietary ToolingImageBounding Box

Developed a neural network to identify and classify cities from satellite images by segmenting large images into smaller partitions for model training. Used LabelImg for annotating key features in images and trained the model with TensorFlow to recognize and compare urban areas. This project enabled advanced geospatial analysis by achieving high accuracy in city identification, optimizing workflows in urban planning.

2022 - 2022

Education

U

Universidad Distrital Francisco José de Caldas

Bachelor's in Cadastral and Geodetic Engineering, Ingeniería Catastral y Geodesia

Bachelor's in Cadastral and Geodetic Engineering

2018 - 2025

Work History

I

INSTITUTO GEOGRAFICO AGUSTIN CODAZZI

Full Stack Developer

Bogotá

2024 - 2024

I

INSTITUTO GEOGRAFICO AGUSTIN CODAZZI

Researcher at the Subdirectorate of Cartography and Geodesy.

BOGOTA

2023 - 2023