Ferney Andres Castaño Estrada

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Top Label Types

Classification

Text Generation

Text Summarization

Diagnosis

Computer Programming Coding

Freelancer Overview

I am a data science enthusiast with hands-on experience in AI training data, data labeling, and annotation, particularly through my work with Outlier.AI, where I ensured high-quality, accurate datasets for machine learning model training. My background in geological engineering and consultancy has given me a strong foundation in managing complex data pipelines, quality control, and integrating AI-driven solutions across domains such as geoscience, biomedical engineering, and mining. I am proficient in Python, R, SQL, and key ML libraries like TensorFlow, Keras, and Scikit-learn, and have practical experience with data visualization, geospatial analysis, and cloud deployment. My projects include developing multimodal chatbots, predictive models for medical diagnostics, and custom data analysis tools, all with a focus on clean, reliable, and well-annotated data to power robust AI systems.

Labeling Experience

Outlier.AI Tier 3 Software Engineer for AI Training

Scale AIComputer Code ProgrammingClassificationText Generation

I worked as a Tier 3 (Expert) Software Engineer contributor focusing on training LLMs in Python, R, and STEM subjects. The scope included RLHF (Reinforcement Learning from Human Feedback) and SFT (Supervised Fine-Tuning). My specific tasks involved evaluating AI-generated code for syntax errors, logic, and efficiency, as well as rewriting responses to meet strict "Gold Standard" quality guidelines. I also created original, complex prompts requiring multi-step reasoning to stress-test the models. I consistently maintained a high quality score, ensuring code executability and factual accuracy.

I worked as a Tier 3 (Expert) Software Engineer contributor focusing on training LLMs in Python, R, and STEM subjects. The scope included RLHF (Reinforcement Learning from Human Feedback) and SFT (Supervised Fine-Tuning). My specific tasks involved evaluating AI-generated code for syntax errors, logic, and efficiency, as well as rewriting responses to meet strict "Gold Standard" quality guidelines. I also created original, complex prompts requiring multi-step reasoning to stress-test the models. I consistently maintained a high quality score, ensuring code executability and factual accuracy.

2024 - 2025

Education

Work History