Not B.

Key Skills

Software

Other

Top Subject Matter

Custom AI assistant and chatbot automation

AI image classification pipeline for door security

Top Data Types

Text

Image

Top Task Types

Text Generation

Classification

Entity (NER) Classification

Freelancer Overview

I don't have formal experience in data labeling or AI training, but I've been building on top of AI models for a while now. I've worked with LLM APIs like Mistral, Groq, and OpenRouter - I know how prompts work, how models respond, and what good vs bad output looks like. I built a full voice assistant where the AI had to return structured JSON with specific actions, emotions, and speech — so I've spent a lot of time thinking about how to get consistent, accurate responses out of a model. I also designed a concept system for thyroid cancer recurrence prediction using ultrasound imaging and an AI classification pipeline - so I understand how AI is applied in high-stakes situations where accuracy actually matters. I'm 16, self-taught in basically everything, and I pick things up fast. I might not have a CV full of labeling jobs but I understand AI better than most people applying for this kind of work - and I'll put in the effort to prove it.

Labeling Experience

AI Model Output Evaluation & Prompt-Response Quality Testing

TextClassification

Developed and tested a multi-provider AI assistant (JARVIS) that required evaluating LLM outputs for accuracy, structure, and consistency. Assessed model responses across providers including Mistral, Groq, and OpenRouter — identifying failures, prompt weaknesses, and incorrect outputs. Designed a strict JSON response schema and iteratively refined prompts to improve model reliability. Also designed an AI classification pipeline for a thyroid cancer recurrence prediction concept, requiring careful thinking around model accuracy and output quality in a high-stakes medical context.

2024 - Present

Education

Work History