Abdullah S.

Key Skills

Software

Label Studio

Top Subject Matter

AI training for LLM output evaluation and Web3/finance content quality

Prompt engineering for RLHF/SFT-style AI training

Domain-expert review and entity-centric research labeling for Web3/DeFi

Top Data Types

Text

Document

Top Task Types

Prompt + Response Writing (SFT)

Data Collection

RLHF

Evaluation/Rating

Text Summarization

Freelancer Overview

Freelance Web3 Content Creator & DeFi Analyst (AI/content quality evaluation and alignment-style review) — 2024–Present. Brings 3+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Don't disclose, Other, and Internal. Education includes Bachelor of Science, Abubakar Tafawa Balewa University (ATBU) (2022). AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Evaluation, Rating, and Prompt + Response Writing (SFT).

Labeling Experience

Founder & Developer — TrenchReads On-Chain Token Rugcheck Tool (2025–Present)

Data Collection

Built Web3 tooling and research outputs that support dataset creation for AI training. Integrated third-party APIs and created a workflow that collects structured token safety signals suitable for subsequent text annotation or evaluation. Produced repeatable data retrieval behavior that can be used as evidence sources for domain-expert labeling. • Integrated DexScreener and GoPlus Security data via a local CORS proxy. • Developed a standalone rugcheck tool to retrieve token safety signals. • Planned/implemented bot + SaaS workflow to automate data collection. • Packaged evidence-based token research outputs in an AI-ready format.

2025 - Present

Domain expert review via smart contract and on-chain research (2024–Present)

Don't discloseText

Performed domain-expert review and structured analysis of Web3/DeFi content, supporting downstream annotation and labeling use cases. Identified and organized key entities related to tokens, protocols, wallets, and contract functions to improve dataset usability. Translated research findings into consistent textual signals that can be reused for instruction-response quality scoring. • Extracted salient entities from smart-contract and on-chain research narratives. • Organized token/protocol/function references into structured explanations. • Reviewed factual consistency of domain claims in AI-ready outputs. • Converted research notes into label-consumable text evidence.

2024 - Present

AI prompting and content/prompt preparation work (self-directed, 2024–Present)

OtherTextPrompt + Response Writing (SFT)

Crafted high-quality prompts for domain-specific queries across Web3, finance, and general topics to support instruction following and reasoning. Wrote and iterated prompt sets consistent with RLHF-style preference workflows by varying instructions and response constraints. Produced prompt-driven content that can be used as training material for instruction-response behavior in LLMs. • Built prompts for instruction-following and domain-specific question answering. • Iterated prompts to elicit clearer, more accurate responses. • Incorporated reasoning and creative writing constraints in prompt variants. • Prepared structured text outputs suitable for SFT-style pipelines.

2024 - Present

Freelance Web3 Content Creator & DeFi Analyst (AI/content quality evaluation and alignment-style review) — 2024–Present

Don't discloseText

Provided AI output quality evaluation for written content in Web3 and finance contexts, focusing on clarity, factual accuracy, and helpfulness. Flagged hallucinations and low-quality or unsafe responses using structured personal rubrics aligned to alignment and safety criteria. Used preference-style comparisons to support selection of higher-quality responses for end-user usefulness. • Ranked multiple model responses by quality and safety. • Identified hallucinations and misinformation in AI-generated text. • Applied quality dimensions such as tone, coherence, and instruction adherence. • Produced evidence-based research outputs that reflect evaluation findings.

2024 - Present

Education

Work History