Samuel O.

Key Skills

Software

Mercor

Micro1

Mindrift

Scale AI

Remotasks

Top Subject Matter

LLM output evaluation and RLHF

Model alignment evaluation and safety-focused rating

LLM code evaluation

Top Data Types

Image

Video

Audio

Top Task Types

Segmentation

Text Generation

Object Detection

Bounding Box

Text Summarization

Transcription

Evaluation/Rating

Fine-tuning

Freelancer Overview

AI Trainer — Mercor. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Mercor, Micro1, and Turing. Education includes Bachelor of Science, Idaho State University (2021). AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Evaluation, Rating, and Computer Programming.

Labeling Experience

AI Trainer — Mercor

MercorText

Provided expert evaluations and ratings for LLM outputs to improve accuracy, reasoning quality, and code correctness. Performed RLHF-related evaluation tasks including comparative response assessment. Identified failure patterns such as hallucinations and edge cases to guide iterative model refinement. • Accuracy and reasoning quality scoring • Code correctness checks for generated solutions • Comparative response rating for RLHF • Reporting of hallucinations and reasoning failures

2024 - Present

AI Trainer — Micro1

Micro1Text

Evaluated and ranked AI-generated responses for coherence, relevance, safety, and instruction-following. Designed targeted prompts to probe model behavior across different domains and reasoning styles. Contributed structured annotation data to model alignment pipelines to improve output quality metrics. • Response ranking and rubric-based evaluation • Prompt design for behavior testing • Structured annotation for alignment pipelines • Bias and failure-mode identification

2023 - 2024

AI Trainer — Turing

Reviewed and annotated AI-generated code to support LLM training pipelines. Applied software engineering judgment to identify logic errors, inefficiencies, and poor-quality outputs. Participated in red-teaming to uncover model vulnerabilities and boundary-condition failures. • Code review and annotation for training • Logic-error and efficiency detection • Red-teaming and vulnerability surfacing • Structured evaluation across 500+ tasks

2021 - 2023

Education

Work History