Tabiro A.

Key Skills

Software

Scale AI

Appen

Remotasks

Surge AI

Label Studio

Other

Don't disclose

Top Subject Matter

NLP dataset annotation for machine learning and LLM fine-tuning (text, intent, safety/moderation)

Speech and conversational audio annotation for ML/NLP datasets

RLHF evaluation and preference labeling for LLM fine-tuning

Top Data Types

Text

Audio

Document

Top Task Types

Entity (NER) Classification

Transcription

RLHF

Prompt + Response Writing (SFT)

Classification

Freelancer Overview

Freelance AI Data Annotator & Labeller (Remote) — Multiple AI/ML Clients. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Scale AI, Appen, and Remotasks. Education includes Bachelor of Arts, University of Nairobi (2020) and N/A, Coursera / Stanford University (2023). AI-training focus includes data types such as Text and Audio and labeling workflows including Entity (NER) Classification, Transcription, and RLHF.

Labeling Experience

Freelance AI Data Annotator & Labeller (Remote) — Multiple AI/ML Clients

OtherTextPrompt + Response Writing (SFT)

Contributed to instruction-following dataset creation to fine-tune large language models. Wrote, rewrote, and evaluated prompts and responses to produce high-quality supervised training examples. Applied guideline-driven quality checks to ensure instructions and outputs align with desired behavior and policy constraints. • Authored and revised prompts and responses for instruction datasets • Evaluated training examples for consistency, correctness, and suitability • Supported SFT-ready data generation for LLM fine-tuning pipelines • Escalated edge cases and ambiguous scenarios for guideline clarification

2022 - Present

Freelance AI Data Annotator & Labeller (Remote) — Multiple AI/ML Clients

Don't discloseTextRLHF

Performed RLHF annotation tasks by ranking and rating AI model responses across quality, helpfulness, factual accuracy, and safety dimensions. Applied rating rubrics to distinguish preference outcomes and to ensure consistent evaluation across annotators and projects. Flagged unsafe or noncompliant responses to support moderation-aware model training. • Ranked and rated responses for helpfulness and preference outcomes • Evaluated factual accuracy and safety/harmlessness per rubric criteria • Reviewed and corrected model outputs to strengthen human feedback loops • Maintained quality through benchmarked processes and inter-annotator agreement

2022 - Present

Freelance AI Data Annotator & Labeller (Remote) — Multiple AI/ML Clients

OtherAudioTranscription

Annotated conversational audio data to support NLP model training and evaluation workflows. Produced verbatim transcriptions and corrected machine-generated speech outputs using human review to improve downstream dataset quality. Ensured consistent utterance boundaries and alignment for reliability across high-volume annotation tasks. • Transcribed and reviewed conversational audio datasets (including ASR transcript correction) • Applied guideline-based checks for clarity, boundaries, and consistency • Coordinated with project-specific safety and content standards during audio review • Supported human-in-the-loop RLHF preparation by assessing response quality attributes

2022 - Present

Freelance AI Data Annotator & Labeller (Remote) — Multiple AI/ML Clients

Scale AITextEntity (NER) Classification

Freelance AI Data Annotator and Labeller supporting NLP and LLM training pipelines by applying detailed annotation guidelines across multiple projects. Annotated text data for named entity recognition, sentiment and toxicity-related classification, and coreference resolution while maintaining high accuracy and consistency. Performed human-in-the-loop reviews of AI outputs to improve quality and safety for downstream model performance. • Labeled NER, sentiment, toxicity/policy risk, and coreference links per project guidelines • Reviewed and corrected ASR transcripts, chatbot responses, and summarizations for improved model behavior • Flagged ambiguous, harmful, or out-of-scope content according to moderation and safety rules • Maintained >98% annotation accuracy with inter-annotator agreement protocols

2022 - Present

Communications Associate — Nairobi Media & Research Consultancy

OtherTextClassification

Supported research-oriented data collection and thematic labeling for qualitative spoken content. Transcribed and coded interview and focus-group data using thematic labeling frameworks analogous to annotation pipelines in NLP research. Ensured consistency across large volumes of text and audio content by following coding guidelines and refining methodologies with senior researchers. • Collected qualitative interview and focus-group data for analysis workflows • Transcribed, coded, and categorized spoken content using thematic frameworks • Maintained inter-rater reliability through adherence to coding guidelines • Produced structured summaries and reports to inform downstream researchers

2021 - 2022

Education

Work History