Open to opportunities

Patrick Mutua

@patrickmutua

Message

Highly analytical AI Content Trainer improving multilingual LLMs through RLHF, SFT, and rigorous evaluation.

Kenya

Message

What I'm looking for

I’m looking to join a team that values rigorous evaluation—pairwise comparisons, rubrics, and root-cause analysis—to harden multilingual LLM outputs. I want work that blends cultural and linguistic insight with impactful, measurable model improvements.

I’m a highly analytical AI Content Trainer and Data Annotator who optimizes advanced language models by identifying multilingual linguistic failure modes and isolating reproducible error traces. In my work with Outlier & Prolific, I pressure-test LLM behavior using prompt engineering, pairwise comparisons, and rubric-based scoring across logic, factual accuracy, syntax, and semantics—so engineering teams can harden model reasoning with evidence-backed findings.

Beyond model evaluation, I bring cultural consultancy and cross-linguistic analysis that improves dataset alignment for diverse demographics. As a freelance Cultural Consultant since 2021, I research localized language trends, advise on cultural sensitivity and appropriateness for international datasets, and help teams collaborate across remote environments; I’ve also delivered high-quality labeling and annotations with strict QA metrics and transcribed complex audio to preserve exact textual representation.

Experience

Work history, roles, and key accomplishments

Current

AI Content Trainer & Annotator

Current

Outlier & Prolific

Nov 2025 - Present (7 months)

Evaluated and stress-tested advanced LLMs using prompt engineering, pairwise comparisons, and rubric-based scoring to identify multilingual failure modes. Documented reproducible error traces across logic, factual accuracy, syntax, and semantics to support model hardening.

LLM Evaluation Prompt Engineering Pairwise Comparisons Error Trace Documentation Root Cause Analysis

Current

Cultural Consultant

Current

Freelance

Feb 2021 - Present (5 years 4 months)

Researched cultural contexts and localized language trends to ensure contextually accurate dataset inputs. Advised engineering and content teams on cultural sensitivity and built alignment-focused content strategies for diverse demographics.

Cultural Consultancy Localization Cross Cultural Communication

Data Annotation Specialist

Data Annotation Tech

Sep 2025 - Nov 2025 (2 months)

Assessed language model outputs for factual soundness and linguistic precision across targeted language groups. Analyzed logical flow, syntactic structure, and semantic accuracy, and documented edge cases and failure modes to refine training data.

Data Annotation Edge Case Reporting

Data Labeler & Annotator

Atlas Capture

Mar 2025 - Sep 2025 (6 months)

Performed high-quality data labeling and annotation to ensure clean, high-fidelity inputs for machine learning models. Maintained strict adherence to quality assurance metrics and achieved top-tier accuracy ratings.

Data Labeling Quality Assurance Annotations Data Cleaning High Fidelity Inputs

Audio Transcriptionist

Crowd Gen

Jan 2025 - Mar 2025 (2 months)

Transcribed complex audio with attention to phonology, morphology, and syntax to produce exact textual representations. Ensured transcription quality through detailed linguistic interpretation of challenging speech content.

Audio Transcription Morphology Syntax Quality Control