Ouko P.

Key Skills

Software

Scale AI

Telus

Appen

Other

iMerit

Top Subject Matter

Autonomous vehicle perception (lidar point cloud)

Search quality and voice assistant training data

LLM training (RLHF) and content moderation QA

Top Data Types

3D Sensor

Audio

Text

Image

Document

Top Task Types

Segmentation

Entity (NER) Classification

RLHF

Freelancer Overview

AI Data Trainer – Freelance Contributor (Scale AI / Telus AI / Silencio AI). Professional background includes roles such as AI Data Trainer. Core strengths include Scale AI, Telus, and Appen. Education includes Bachelor of Education, Kenyatta University (2025). AI-training focus includes data types such as 3D Sensor, Audio, and Text and labeling workflows including Segmentation, Entity (NER) Classification, and RLHF.

Labeling Experience

AI Data Trainer – Freelance Contributor (Scale AI / Telus AI / Silencio AI)

Scale AI3D SensorSegmentation

Completed large-scale annotations for autonomous vehicle perception on 3D lidar point clouds, including relevant object delineation for downstream model training and validation. Performed detailed labeling of spatial regions and ensured consistency with project guidelines for reliable perception outputs. Supported data pipeline QA through careful review of labeled artifacts. • Labeled lidar point cloud data for autonomous vehicle perception models • Produced annotations supporting 3D understanding tasks for training • Maintained accuracy and guideline adherence for top contributor status • Contributed to high-volume dataset expansion across projects

2026 - 2024

Multilingual Dialogue Dataset — NLP Annotation Lead (Independent Contract)

OtherText

Led annotation for a multilingual dialogue dataset by labeling discourse acts, coreference chains, and sentiment for cross-lingual conversational AI research. Created high-quality instruction–response pairs for instruction tuning workflows using a large language model. Ensured labels were consistent and usable for downstream fine-tuning and evaluation. • Annotated 15,000 English–Swahili dialogue turns with discourse/sentiment labels • Labeled coreference chains to support dialogue understanding • Wrote 500 instruction–response pairs for model fine-tuning • Supported multilingual conversational AI dataset contributions

2023 - 2023

Data Entry & Quality Assurance Associate (Telus International AI)

TelusAudioEntity (NER) Classification

Evaluated and rated search quality outputs to improve relevance, intent match, and information correctness for a global search quality initiative. Annotated intent and entity information within voice-assistant style utterances to support training data creation at scale. Identified and escalated systematic annotation defects to reduce project-wide error rates. • Rated web search results for relevance and intent alignment • Annotated intent and entity labels across 10,000+ utterances/month • Flagged systematic errors to reduce defects over multiple quarters • Performed quality checks aligned to search quality standards

2023 - 2022

Medical Image Annotation — Bounding Box & Segmentation (iMerit Contract)

iMeritSegmentation

Labeled medical imaging data under clinical supervision by creating anatomical bounding boxes and pixel-level segmentation masks. Followed HIPAA-compliant data handling procedures throughout the labeling lifecycle to preserve patient privacy. Produced structured annotations suitable for downstream radiology perception model development. • Labeled 8,000 radiology images with bounding boxes and segmentation masks • Created pixel-level segmentation under expert supervision • Followed HIPAA-compliant data handling protocols • Delivered clinical-grade annotations for model training

2022 - 2022

Senior Data Annotation Specialist (Appen)

AppenTextRLHF

Annotated and rated AI training data spanning text, image, and audio tasks while maintaining high quality scores and consistent guideline compliance. Completed RLHF-style work by rating and ranking LLM-generated responses along helpfulness, harmlessness, and honesty to inform model fine-tuning. Also reviewed policy and content moderation datasets and acted as a peer reviewer to improve annotation reliability. • Rated and ranked LLM responses for RLHF (helpfulness/harmlessness/honesty) • Reviewed content moderation labels for hate speech, misinformation, and NSFW • Peer-reviewed and audited 200+ annotations/week from junior contributors • Calibrated guidelines and maintained IAA (Cohen’s Kappa >0.85)

2022 - 2022

Education

Work History