Johnson O.

Key Skills

Software

Micro1

Mercor

Mindrift

OneForma

Labelbox

Humanatic

Mighty AI

Prodigy

Redbrick AI

Remotasks

Roboflow

Sama

V7 Labs

Trilldata Technologies

VoTT

Telus

Toloka

Tagtog

Surge AI

SuperAnnotate

Snorkel AI

Sloth

Scale AI

Supervisely

Playment

OpenCV AI Kit (OAK)

Anno-Mage

Argilla

Appen

Axiom AI

Clickworker

CloudFactory

CrowdFlower

CrowdSource

Data Annotation Tech

Datatroniq

Dataloop

Datumbox

Encord

Google Cloud Vertex AI

Hasty

HiveMind

Top Subject Matter

LLM Agent Training

Conversational AI

Task Automation

Top Data Types

Computer Code Programming

3D Sensor

Geospatial Tiled Imagery

Top Task Types

Prompt + Response Writing (SFT)

Classification

Function Calling

Bounding Box

Object Detection

Text Generation

Text Summarization

RLHF

Fine-tuning

Transcription

Computer Programming/Coding

Polygon

Segmentation

Entity (NER) Classification

Polyline

Point/Key Point

Cuboid

Question Answering

Red Teaming

Evaluation/Rating

Data Collection

Freelancer Overview

AI Software Engineer – Morningtrain. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal and Proprietary Tooling. Education includes a Bachelor of Science from Obafemi Awolowo University (2023) and a Senior Secondary Certificate Examination from Libration Academy (2014). AI-training focus includes data types such as Text and labelling workflows, including Prompt + Response Writing (SFT) and Classification.

Labeling Experience

AI Software Engineer – Morningtrain

TextPrompt + Response Writing (SFT)

As an AI Software Engineer at Morningtrain, I was responsible for designing and generating complex conversational datasets for LLM agent training. I engineered evaluation data and simulated realistic multi-turn assistant-user dialogue flows to benchmark model performance. My role involved quality assurance, iterative dataset refinement, and ensuring high compliance with annotation guidelines. • Generated 1,200+ multi-turn conversation samples covering scheduling, email, maps, and automation domains • Built function-calling simulation workflows using structured JSON data • Created evaluation datasets with 300+ test cases to benchmark reasoning accuracy • Improved dataset acceptance rates through iterative data refinement

2024 - 2026

Computer Vision Data Annotator

3D SensorBounding Box

Annotated 5,000+ images and video frames using bounding boxes and segmentation Labeled objects for autonomous systems, robotics, and real-time vision AI models Worked on datasets aligned with 3D, video, and multimodal annotation pipelines Ensured consistency across large datasets using defined taxonomies Performed QA checks to improve model training accuracy Delivered datasets used in smart surveillance, retail analytics, and industrial AI systems

2025 - 2025

NLP Data Annotation Specialist

TextClassification

Annotated 8,000+ data points for intent classification, NER, and sentiment analysis Labeled LLM preference data and response comparisons for model alignment (RLHF/DPO) Structured multi-turn conversations for chatbot and AI agent training Tagged entities and contextual signals for improved reasoning and response quality Validated datasets to meet production-level AI standards Supported training of enterprise AI assistants and agent-based systems

2024 - 2024

AI Content Moderation Annotator

ImageClassification

Labelled 12,000+ text and image samples across hate speech, harassment, and misinformation categories Applied policy guidelines to resolve edge cases and maintain consistency Conducted QA reviews and corrected mislabels across batches Maintained 98%+ annotation accuracy Delivered structured datasets used to train moderation models

2024 - 2024

LLM Data Trainer – Yorubatrainers

TextPrompt + Response Writing (SFT)

As an LLM Data Trainer at Yorubatrainers, I created annotated training samples for instruction-following LLMs with a focus on tool usage and reasoning. I simulated dual-role dialogues and generated high-fidelity datasets to support model alignment and consistency. My responsibilities included quality assurance, scenario creation, and correction of model output errors. • Produced 900+ annotated samples for tool-following and reasoning • Simulated both user and assistant roles for realistic data generation • Identified and corrected over 250 instances of tool misuse or hallucinations • Increased dataset consistency score through scenario and template development

2023 - 2024

Show more

Education

Work History