Marc M.

Key Skills

Software

Mercor

Other

Top Subject Matter

LLM evaluation

red teaming

and supervised fine-tuning dataset/prompt & rubric creation for AI models

Top Data Types

Text

Document

Top Task Types

Red Teaming

Fine-tuning

Freelancer Overview

Mercor (Remote) — Senior Writer/Reviewer, Supervised Fine Tuning (SFT) / Critical Failure Induction. Brings 17+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Mercor and Other. Education includes Bachelor of Science, Penn State University (2010) and Study Abroad (Non-Degree), Palazzo Rucellai (2009). AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Red Teaming and Fine-tuning.

Labeling Experience

Handshake AI

VideoTranscription

Perform transcription of all audio and video in various clips

2026 - Present

Mercor (Remote) — Project Plato/Horizon/Labyrinth (SFT & agency reliability work)

OtherFine-tuning

Fine-tuning protocol development and evaluation-oriented training work for LLM agents aimed at improving autonomy and decision accuracy. The work included designing prompting strategies and training data/datasets to enforce formatting and behavioral constraints while supporting cross-platform execution. Iterative failure analysis was used to refine decision-making and improve reliability for financial and administrative tasks. • Developed fine-tuning protocols and complex prompting strategies for LLM agency and autonomous task execution across software tools. • Engineered end-to-end workflows for LLM navigation/manipulation of external software interfaces and reduced manual data entry/report generation. • Created datasets and constraints to enforce complex markdown/professional formatting without degrading conversational flow. • Refined model logic through iterative failure analysis to ensure high-accuracy outputs in sensitive financial/admin tasks.

2025 - Present

Mercor (Remote) — Senior Writer/Reviewer, Supervised Fine Tuning (SFT) / Critical Failure Induction

MercorTextRed Teaming

Supervised fine-tuning (SFT) support and rigorous evaluation work for large language models, centered on inducing failures and grading output quality. The role involved creating prompt sets and rubrics to assess reasoning, numerical accuracy, domain expertise, compliance, and failure modes. Outputs were systematically tested via adversarial and edge-case prompt construction to surface issues such as ambiguity, leakage, and scoring weaknesses. • Authored high-difficulty prompts for evaluation of LLM reasoning, numerical accuracy, and domain expertise. • Developed formal grading rubrics to score correctness, reasoning quality, compliance, and failure modes. • Performed prompt and rubric reviews while inducing and analyzing critical model failures. • Identified ambiguity, leakage, and rubric/scoring weaknesses to improve reliability.

2025 - Present

Education

Work History