Jonathan Martinez

Key Skills

Software

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Text

Video

Top Task Types

Action Recognition

Audio Recording

Classification

Prompt Response Writing SFT

Question Answering

Freelancer Overview

My experience in data labeling/AI training begun with Data Annotation Tech. I was tasked with working on several different projects with varying tasks. My first few tasks consisted of working on a chatbot project named "Raven," where I was instructed to create and rate criteria that the AI model would follow for the purposes of response generation. I would have to create criteria based on the bot's system prompt and the user's prompt itself. I begun working on other projects shortly after. A project named "Achilles" took up most of my time where I had to rate two responses to a user prompt. Each response was graded on truthfulness, grammar and overall helpfulness. I also worked on Project Bluebird, which was a series of tasks working with Gemini and ChatGPT, rating their performance on screen-sharing and video-sharing prompts.

Labeling Experience

Bluebird

Data Annotation TechVideoQuestion AnsweringEvaluation Rating

Bluebird involved using Gemini and ChatGPT's Screen-sharing and Video-sharing features. I was instructed with asking questions about what is being shown in the video and having a conversation afterwards. The AI responses would then be judged on several factors. I was tasked with then choosing which AI model was better for the given task.

Bluebird involved using Gemini and ChatGPT's Screen-sharing and Video-sharing features. I was instructed with asking questions about what is being shown in the video and having a conversation afterwards. The AI responses would then be judged on several factors. I was tasked with then choosing which AI model was better for the given task.

2024 - 2024

Raven

Data Annotation TechTextText GenerationPrompt Response Writing SFT

Raven consisted of creating criteria or rules that the AI would follow to generate it's responses.

Raven consisted of creating criteria or rules that the AI would follow to generate it's responses.

2024 - 2024

Achilles

Data Annotation TechTextEvaluation RatingPrompt Response Writing SFT

The Achilles project consisted of working on rating AI responses and ensuring they were factual. The project provided a prompt and two responses. One response was to be chosen as the better of the two.

The Achilles project consisted of working on rating AI responses and ensuring they were factual. The project provided a prompt and two responses. One response was to be chosen as the better of the two.

2024 - 2024

Education

Work History