Rama Marín

Key Skills

Software

Clickworker

Remotasks

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Evaluation Rating

Object Detection

Prompt Response Writing SFT

Question Answering

Text Generation

Freelancer Overview

My background includes proven experience in both AI training data quality control and advanced data analysis, providing me with a comprehensive skill set for managing, refining, and validating complex information. As an AI trainer/data quality analyst, I specialized in performing detailed data annotation and scoring tasks on Spanish AI outputs. This role required attention to detail, providing constructive written feedback, and ensuring strict adherence to guidelines to improve model quality, accuracy, and reliability. I have experience in various classification tasks, including entity and sentiment tagging. Complementing this focus on quality, my experience as a data analyst has equipped me with strong technical skills in Python, SQL, and AWS. I am adept at building robust data pipelines, having designed solutions to ingest API data, load it into PostgreSQL databases, and prepare it for visualization. This dual perspective, combining rigorous quality control of AI systems with practical data engineering skills.

Labeling Experience

Source Verification for LLM Factual Accuracy

Scale AITextText GenerationEvaluation Rating

Executed specialized Quality Control (QC) tasks focused on validating the factual accuracy of AI-generated text. The scope involved meticulously analyzing every factual data point provided by the LLM. Responsibilities included conducting targeted online research to locate at least one official and reliable source for verification. Successfully detected and documented AI hallucinations, and classified the text as either correct or factually inaccurate based on external evidence. All evaluations required precise citation logging (URL pasting) to justify the final assessment and enhance the model's reliability metrics.

Executed specialized Quality Control (QC) tasks focused on validating the factual accuracy of AI-generated text. The scope involved meticulously analyzing every factual data point provided by the LLM. Responsibilities included conducting targeted online research to locate at least one official and reliable source for verification. Successfully detected and documented AI hallucinations, and classified the text as either correct or factually inaccurate based on external evidence. All evaluations required precise citation logging (URL pasting) to justify the final assessment and enhance the model's reliability metrics.

2024 - 2024

Prompt Engineering & Advanced AI Response Generation

Scale AITextComputer Programming CodingPrompt Response Writing SFT

I executed an advanced text generation project focused on prompt engineering and quality benchmarking. The tasks involved generating highly specific prompts based on detailed project guidelines to test the capabilities of the LLM. The primary responsibility was to simulate a superior AI response, establishing a high-quality human benchmark. This required generating comprehensive, factual, and natural-sounding text that consistently outperformed the base model's results. This work directly contributed to refining the model's fluency, creativity, and compliance with complex contextual instructions.

I executed an advanced text generation project focused on prompt engineering and quality benchmarking. The tasks involved generating highly specific prompts based on detailed project guidelines to test the capabilities of the LLM. The primary responsibility was to simulate a superior AI response, establishing a high-quality human benchmark. This required generating comprehensive, factual, and natural-sounding text that consistently outperformed the base model's results. This work directly contributed to refining the model's fluency, creativity, and compliance with complex contextual instructions.

2024 - 2024

Education

Work History