Brodsky Michael Alan

Key Skills

Software

CVAT

Data Annotation Tech

Figure Eight

Labelbox

Label Studio

Mercor

Mindrift

OpenCV AI Kit (OAK)

Snorkel AI

Toloka

VoTT

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Classification

Evaluation Rating

Prompt Response Writing SFT

RLHF

Text Summarization

Freelancer Overview

I am an analytical and detail-oriented AI Trainer with a background in physics, English and mathematics, specializing in evaluating model reasoning, accuracy, and problem-solving performance. Over the past two years, I have worked on diverse data labeling and AI training projects focused on assessing logical coherence, validating scientific responses, and developing complex physics-based tasks to refine AI comprehension and reliability. My expertise lies in structured data annotation, model evaluation, and the design of domain-specific prompts that enhance model understanding. Combining a rigorous scientific approach with proficiency in Python, MATLAB, and AI evaluation tools, I consistently deliver high-quality, precise labeling outputs that strengthen data integrity and model performance.

Labeling Experience

AI Physics Model Evaluation Study

Data Annotation TechText

In this project, I evaluated AI-generated physics solutions across mechanics, thermodynamics, and electromagnetism. I developed over 150 domain-specific prompts and assessed model responses for logical accuracy, completeness, and reasoning quality. Feedback was documented and used to guide fine-tuning iterations, improving model correctness by over 25%. The work involved detailed annotation, structured scoring, and consistency checks to ensure high-quality evaluation data.

In this project, I evaluated AI-generated physics solutions across mechanics, thermodynamics, and electromagnetism. I developed over 150 domain-specific prompts and assessed model responses for logical accuracy, completeness, and reasoning quality. Feedback was documented and used to guide fine-tuning iterations, improving model correctness by over 25%. The work involved detailed annotation, structured scoring, and consistency checks to ensure high-quality evaluation data.

2018 - 2025

General Reasoning and Text Generation Evaluation

MindriftTextRLHFEvaluation Rating

Evaluated and annotated AI-generated text for correctness, clarity, and human-likeness across diverse topics including general reasoning, education, and everyday dialogue. Tasks included writing prompts, assessing model outputs, and scoring them on factual accuracy, contextual understanding, and ethical compliance. Contributed to data used in fine-tuning conversational AI systems and improving safety alignment. Maintained high-quality annotation standards and followed strict QA protocols to ensure dataset consistency.

Evaluated and annotated AI-generated text for correctness, clarity, and human-likeness across diverse topics including general reasoning, education, and everyday dialogue. Tasks included writing prompts, assessing model outputs, and scoring them on factual accuracy, contextual understanding, and ethical compliance. Contributed to data used in fine-tuning conversational AI systems and improving safety alignment. Maintained high-quality annotation standards and followed strict QA protocols to ensure dataset consistency.

2024

Education

Work History