Peter O.

Key Skills

Software

Other

Top Subject Matter

STEM (mechanical engineering, petroleum systems, process safety, statistical analysis, data science)

Research documentation quality assurance and evidence-based reasoning

Government statistics

Top Data Types

Text

Document

Top Task Types

RLHF

Text Generation

Data Collection

Freelancer Overview

AI Trainer & Data Annotation Specialist (Freelance/Independent). Brings 8+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Other. Education includes Doctor of Philosophy, University of Portsmouth (2023) and Master of Science, University of Portsmouth. AI-training focus includes data types such as Text and Document and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

AI Trainer & Data Annotation Specialist (Freelance/Independent)

OtherTextRLHF

Provided AI response evaluation and rating for LLM training programs using five quality dimensions: instruction following, factual accuracy, reasoning quality, clarity, and harmlessness. Produced detailed written rationales for each evaluation decision to support model improvement and reduce errors. Calibrated confidence to evidence to ensure outputs follow appropriate epistemic standards rather than false certainty.• Rated AI responses across RLHF-style criteria with consistently high accuracy.• Performed STEM-focused annotation requiring subject-matter expertise in mechanical engineering and petroleum systems.• Wrote and refined prompts plus ideal responses for RLHF training datasets using structured reasoning frameworks.• Identified hallucinations, factual errors, and logical inconsistencies, flagging bias and uncertainty when applicable.

2025 - Present

Research Manager (Contractor), University of Southampton

OtherDocument

Produced and evaluated large volumes of structured, evidence-based written outputs analogous to AI training data annotation. Designed standardized assessment instruments and evaluation frameworks to measure research quality with the same rigor used in AI response rating. Identified inconsistencies, anomalies, and errors in research datasets and produced structured reports for expert and non-expert audiences.• Evaluated written research documentation for factual accuracy, logical coherence, and instruction compliance.• Built evaluation frameworks and assessment instruments for quality measurement.• Performed error/hallucination-style detection by finding dataset inconsistencies and reasoning failures.• Delivered clear, structured written recommendations and reports to multiple stakeholder groups.

2024 - 2025

Statistical Data Analyst (Contractor), Office for National Statistics (ONS)

OtherText

Applied statistical reasoning and quantitative analysis to large-scale government datasets in support of evidence validation and quality compliance. Detected data anomalies, inconsistencies, and error patterns, improving data quality compliance through systematic verification. Translated analytical results into accessible written reports while validating evidence against established standards for reasoning quality and specification adherence.• Performed evidence validation and specification checks analogous to factual accuracy verification in AI evaluation.• Identified recurring data error patterns and anomaly clusters to improve dataset reliability.• Produced clear written analytical reports for non-technical stakeholders with audience-aware clarity.• Assessed where data quality or reasoning failed to meet requirements for annotation/evaluation readiness.

2022 - 2023

PhD Researcher, Mechanical & Design Engineering (University of Portsmouth)

OtherTextText Generation

Conducted seven years of structured, evidence-based research requiring repeated validation of hypotheses, logical reasoning, and factual accuracy verification. Produced peer-reviewed technical publications and conference presentations with calibrated confidence levels applicable to generating ideal AI responses and evaluating AI-generated content quality. Investigated complex failure scenarios in STEM domains, building advanced capability for detecting reasoning errors and logical inconsistencies. • Applied inductive and deductive reasoning to validate hypotheses and identify flaws in analytical frameworks.• Authored technical publications and presentations emphasizing precise, structured writing and calibrated confidence.• Developed ML-based predictive monitoring to understand system behavior, limitations, and failure modes relevant to hallucination/error detection.• Investigated catastrophic and near-miss events to strengthen STEM subject-matter expertise for high-specialist annotation tasks.

2018 - 2023

Senior Researcher (Contract), UK Health & Safety Executive (HSE)

OtherDocument

Evaluated complex evidence against regulatory standards and produced structured written recommendations using evidence assessment and reasoning quality evaluation. Synthesised multiple competing sources to reach defensible, well-calibrated conclusions under scrutiny conditions. Produced structured outputs mirroring the factual accuracy and confidence calibration requirements used in RLHF-style evaluation work. • Assessed evidence quality against regulatory specifications for defensible conclusions.• Verified factual accuracy and reasoning quality in written recommendations.• Performed multi-source synthesis to resolve conflicts while calibrating confidence.• Produced structured written outputs suitable for evaluation and improvement workflows.

2022 - 2022

Show more

Education

Work History