Peter A.

Key Skills

Software

Appen

Other

Don't disclose

Top Subject Matter

LLM evaluation and prompt consistency testing (GPT-4)

Survey data preparation and statistical modeling (AI-notionally adjacent analytics)

Legal Services & Contract Review

Top Data Types

Text

Document

Top Task Types

Classification

RLHF

Freelancer Overview

Statistical Analysis of GPT Output Consistency | Capstone Project (AFIT Statistics Department). Brings 3+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Don't disclose and Other. Education includes Bachelor of Science, Air Force Institute of Technology (AFIT) (2026) and Capstone Project, Air Force Institute of Technology (AFIT) Statistics Department (2025). AI-training focus includes data types such as Text and Document and labeling workflows including Evaluation, Rating, and Computer Programming.

Labeling Experience

AI Trainer – Data Annotation & Evaluation (Part-Time), Outlier.ai (Remote Contract)

OtherTextRLHF

Evaluated and ranked LLM-generated responses across accuracy, helpfulness, safety, and instruction-following for a major AI lab’s RLHF pipeline. Wrote and refined adversarial prompts to probe model reasoning and expose edge-case failure modes. Maintained high annotation agreement using detailed rubric guidelines while contributing structured feedback based on observed model error patterns. • LLM response ranking for RLHF dimensions (accuracy/helpfulness/safety/instruction-following) • Adversarial prompt development for red-teaming and edge-case discovery • Sustained agreement rate above 90% across 150+ tasks per month • Used statistical reasoning to analyze error patterns and improve evaluation rubrics

2024 - Present

Undergraduate Research Assistant – Statistics Dept. (AFIT)

OtherDocument

Supported faculty with data cleaning and exploratory analysis for an undergraduate statistics research project. Worked with a survey dataset to prepare data for downstream analysis and modeling. Communicated results through written reporting and academic presentation. • Cleaned and explored a 10,000-row survey dataset using R and Python. • Built regression models linking demographic variables to academic outcomes. • Summarized insights in a written report for faculty review. • Presented findings at the undergraduate research symposium.

2025 - 2026

Statistical Analysis of GPT Output Consistency | Capstone Project (AFIT Statistics Department)

Don't discloseText

Conducted an academic capstone experiment to quantify GPT-4 response consistency under paraphrased prompts. Used statistical techniques including ANOVA and bootstrap resampling to measure variance in model outputs. Documented experimental setup and results in a written report for assessment and presentation purposes. • Designed prompt sets with semantically equivalent but differently phrased inputs. • Applied ANOVA and bootstrap resampling to quantify response variance. • Evaluated model output consistency as the primary metric. • Produced a detailed written report summarizing findings and methodology.

2025 - 2025

Data Annotation Assistant (Part-Time)

AppenTextClassification

Annotated text datasets for sentiment classification, named-entity recognition, and intent detection for commercial NLP model training. Produced consistent annotations aligned to project guidelines and evaluation benchmarks. Supported dataset quality improvements by identifying and flagging low-quality crowd-sourced outputs. • Sentiment classification annotation on text inputs • Entity recognition and intent detection labeling • QA-focused review and flagging of low-quality annotations • Improved estimated dataset accuracy by ~12%

2024 - 2024

Education

Work History