Ammar Butt

Key Skills

Software

Mercor

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

RLHF

Fine Tuning

Freelancer Overview

I am a qualified Solicitor of England and Wales with over three years of legal expertise and a strong focus on legal AI training data and data annotation. My recent work as a Legal AI Expert involves writing and reviewing complex legal prompts, developing multi-dimensional assessment rubrics, and collaborating with AI research teams to train large language models (LLMs) for statutory interpretation, procedural guidance, and case law analysis. I am proficient in workflow management tools like Airtable and have a proven track record of exceeding quality benchmarks in legal data evaluation. My experience bridges the gap between legal practice and AI, ensuring that annotated data and model outputs meet rigorous professional standards for reasoning, objectivity, and jurisdictional relevance.

Labeling Experience

Expert AI Trainer (Subject Matter Expert)

MercorTextFine Tuning

Developing complex instructions and author 'gold standard' responses to train models on advanced reasoning and technical accuracy. My work focuses on ensuring model outputs adhere to specific formatting, logic, and multi-step constraint satisfaction

Developing complex instructions and author 'gold standard' responses to train models on advanced reasoning and technical accuracy. My work focuses on ensuring model outputs adhere to specific formatting, logic, and multi-step constraint satisfaction

2025

Expert AI Trainer (Subject Matter Expert)

MercorTextRLHF

Developed complex prompts designed to identify performance gaps in model reasoning and instruction-following. Authored detailed, multi-dimensional rubrics to evaluate model outputs and standardise grading criteria. Specialized in testing LLM capabilities by constructing specific prompt-rubric pairings to benchmark model accuracy and logic.

Developed complex prompts designed to identify performance gaps in model reasoning and instruction-following. Authored detailed, multi-dimensional rubrics to evaluate model outputs and standardise grading criteria. Specialized in testing LLM capabilities by constructing specific prompt-rubric pairings to benchmark model accuracy and logic.

2025 - 2025

Education

Work History