For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

Invite to Job
D
Damilola Adewole

Damilola Adewole

Project Manager, Customer Success in Contract Review, Compliance, and Legal Research

Canada flagToronto, Canada
$30.00/hrIntermediateMercorMindriftOneforma

Key Skills

Software

MercorMercor
MindriftMindrift
OneFormaOneForma
AppenAppen
Data Annotation TechData Annotation Tech
RemotasksRemotasks
Scale AIScale AI
HiveMindHiveMind

Top Subject Matter

Legal Services & Contract Review
Regulatory Compliance & Risk Analysis
Legal Research & Document Analysis

Top Data Types

DocumentDocument
TextText
VideoVideo

Top Task Types

Bounding BoxBounding Box
SegmentationSegmentation
Data CollectionData Collection
RLHFRLHF
Text SummarizationText Summarization
Evaluation/RatingEvaluation/Rating
Object DetectionObject Detection
PolygonPolygon
Text GenerationText Generation
Prompt + Response Writing (SFT)Prompt + Response Writing (SFT)
PolylinePolyline
Point/Key PointPoint/Key Point
Question AnsweringQuestion Answering
Red TeamingRed Teaming
Computer Programming/CodingComputer Programming/Coding
TranscriptionTranscription
CuboidCuboid
Fine-tuningFine-tuning
Entity (NER) ClassificationEntity (NER) Classification

Freelancer Overview

My experience in AI training data is rooted in the rigorous demands of high-quality data annotation, reinforcement learning from human feedback (RLHF), and prompt engineering. I specialize in evaluating model outputs for factual accuracy, logical consistency, and safety, ensuring that datasets are curated to align with human intent. My background involves complex instruction-following tasks, multi-turn conversation management, and domain-specific fact-checking, which allows me to effectively bridge the gap between raw data and high-performing machine learning models. What sets me apart is my ability to think critically about the nuance of language, including tone, style, and cultural context. I excel in identifying edge cases and ambiguous prompts that could lead to model hallucinations, and I bring a structured, detail-oriented approach to creating "gold standard" datasets. By combining subject matter expertise with a deep understanding of LLM architectures and limitations, I contribute to the development of safer, more reliable, and more intuitive generative AI systems. Education includes Bachelor of Science, University of Greenwich (2015) and Advanced Diploma, Association of Computer Professionals (2014).

IntermediateEnglish

Labeling Experience

Annotator

ImageRLHF
The project involves acting as an expert judge to evaluate AI-generated responses to complex multimodal requests by comparing two models (Media A and Media B) to see which better follows instructions and produces higher-quality output. The core evaluation is built on four distinct pillars: Instruction Following, which requires strict adherence to every part of a prompt with no partial credit given; Visual Quality, which assesses aesthetics and the preservation of the original image’s structure; AI-Generated Issues, which identifies "Red Flags" like distorted features, melting shapes, or impossible geometry; and Overall Preference, which weighs these factors to determine the most "trustworthy" result. A critical priority in this process is to rank Instruction Following and Realism/Naturalness above general aesthetic quality, ensuring that a "pretty" but incorrect image is not preferred over a more accurate one. Evaluators must use the full Conversation History as the sole source of truth, assessing each category independently while avoiding "Ties" unless no meaningful difference exists

The project involves acting as an expert judge to evaluate AI-generated responses to complex multimodal requests by comparing two models (Media A and Media B) to see which better follows instructions and produces higher-quality output. The core evaluation is built on four distinct pillars: Instruction Following, which requires strict adherence to every part of a prompt with no partial credit given; Visual Quality, which assesses aesthetics and the preservation of the original image’s structure; AI-Generated Issues, which identifies "Red Flags" like distorted features, melting shapes, or impossible geometry; and Overall Preference, which weighs these factors to determine the most "trustworthy" result. A critical priority in this process is to rank Instruction Following and Realism/Naturalness above general aesthetic quality, ensuring that a "pretty" but incorrect image is not preferred over a more accurate one. Evaluators must use the full Conversation History as the sole source of truth, assessing each category independently while avoiding "Ties" unless no meaningful difference exists

2025 - 2026

Annotator

ImageEvaluation Rating
Contributors must first verify that the initial sketch matches the original subject in pose and identity, skipping any tasks where the input is inconsistent or already photographic. The primary objective is to select the most realistic output that mimics an actual camera-captured photo while avoiding images that look like digital renders or CGI. Evaluators are instructed to reject results with unnatural distortions, warped anatomy, or added elements not found in the source sketch. Accuracy is essential, meaning the final selection should preserve key structural features and realistic proportions without artistic exaggeration. Through a structured multi-step process, users iterate and refine generations to ensure the highest level of authenticity and fidelity to the input.

Contributors must first verify that the initial sketch matches the original subject in pose and identity, skipping any tasks where the input is inconsistent or already photographic. The primary objective is to select the most realistic output that mimics an actual camera-captured photo while avoiding images that look like digital renders or CGI. Evaluators are instructed to reject results with unnatural distortions, warped anatomy, or added elements not found in the source sketch. Accuracy is essential, meaning the final selection should preserve key structural features and realistic proportions without artistic exaggeration. Through a structured multi-step process, users iterate and refine generations to ensure the highest level of authenticity and fidelity to the input.

2025 - 2025

Annotator

ImagePrompt Response Writing SFT
Participants must first validate images to confirm they are photorealistic, person-free, and contain suitable movable items rather than permanent fixtures. The workflow involves generating AI outputs, evaluating them for lighting consistency and natural fill, and selecting the most accurate result. Furthermore, users are required to write detailed reconstruction prompts that describe the original furniture's type and placement to facilitate precise digital re-staging

Participants must first validate images to confirm they are photorealistic, person-free, and contain suitable movable items rather than permanent fixtures. The workflow involves generating AI outputs, evaluating them for lighting consistency and natural fill, and selecting the most accurate result. Furthermore, users are required to write detailed reconstruction prompts that describe the original furniture's type and placement to facilitate precise digital re-staging

2025 - 2025

High-Resolution Computer Use Annotator

ImageObject Detection
Participants are required to source high-resolution screenshots of software interfaces and websites that contain at least ten interactable elements. The primary objective is to create precise bounding boxes around every individual icon and button, ensuring no gaps or overlaps exist between the annotations. Strict rules prohibit the use of artificial intelligence for generating prompts or justifications, with violations potentially resulting in platform removal.

Participants are required to source high-resolution screenshots of software interfaces and websites that contain at least ten interactable elements. The primary objective is to create precise bounding boxes around every individual icon and button, ensuring no gaps or overlaps exist between the annotations. Strict rules prohibit the use of artificial intelligence for generating prompts or justifications, with violations potentially resulting in platform removal.

2025 - 2025

Education

U

University of Greenwich

Bachelor of Science, Computing

Bachelor of Science
2012 - 2015
A

Association of Computer Professionals

Advanced Diploma, Computer Science

Advanced Diploma
2014 - 2014

Work History

B

Backbase

Project Manager, Customer Success

Toronto
2024 - 2025
B

BDO Canada

Senior Project Manager, Managed Services

Toronto
2024 - 2024