Peter O.

Key Skills

Software

Appen

Clickworker

CVAT

Dataloop

Datasaur

Doccano

Encord

HiveMind

Labelbox

Label Studio

OneForma

Remotasks

Roboflow

Scale AI

SuperAnnotate

Supervisely

Surge AI

Telus

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Object Detection

Polygon

Segmentation

Tracking

Freelancer Overview

I am an Expert AI Data Annotator and LLM Evaluator with over 3 years of intensive freelance experience, delivering more than 500,000 high-quality annotations across computer vision, video, and language model projects. Specializing in autonomous vehicle datasets (bounding boxes, polygons, semantic segmentation, 3D cuboids, lane marking, traffic signs, pedestrian detection) and large language model evaluation (red-teaming, RLHF, preference ranking, safety & harmful content judging, prompt-response writing). Native in both English and Kiswahili, I bring rare multilingual capability to African-language and code-switching projects. Consistently ranked in the top 5% of contributors on Scale AI, Appen, Remotasks, Surge AI, and other major platforms with 98–100% accuracy scores. Fast, reliable, guideline-following, and experienced with complex multi-stage workflows. Available 40–50 hours/week in EAT timezone with excellent overlap for US and European teams.

Labeling Experience

Video Object Tracking & Action Recognition

CVATVideoBounding BoxClassification

Tracked multiple objects frame-by-frame in long video sequences. Labeled human actions (walking, running, falling, shopping) for security and retail analytics projects. Maintained temporal consistency across 50,000+ frames.

2023

Large Language Model Safety & Preference Evaluation

Scale AITextQuestion AnsweringRLHF

Performed red-teaming, harmful content detection, and RLHF preference labeling on 100,000+ prompts and responses. Ranked model outputs for helpfulness, harmlessness, and honesty. Wrote high-quality SFT prompt-response pairs in English.

2023

Autonomous Vehicle Image & Sensor Annotation

Scale AI3D SensorBounding BoxPolygon

Annotated 200,000+ objects (vehicles, pedestrians, cyclists, traffic signs, lane markings) using 2D bounding boxes, polygons, polylines, and 3D cuboids. Worked on LiDAR-camera fusion datasets for leading AV companies. Consistently scored 99%+ IoU accuracy.

2023

Education

Work History