Chinedu Stanley

Evaluated and ranked AI-generated responses across 500+ prompt pairs per week, improving output quality by identifying hallucinations, factual errors, and reasoning flaws. Wrote high-difficulty prompts and provided rationale-based rankings achieving a 97%+ quality score, and reviewed Python/JavaScript code for logical, edge-case, and security issues.

Designed and customized WordPress and WooCommerce storefronts aligned to client brand identities, creating mobile-responsive experiences with streamlined navigation and checkout flows. Built SEO foundations and improved page speed/Core Web Vitals using image compression, caching, and plugin optimization, and implemented tracking and automations via Meta Pixel, Google Analytics, and email tools.

Created complete brand identity packages for startups and small businesses, including logos, color palettes, typography systems, and brand style guides. Produced social media graphics for 5+ clients and delivered print-ready marketing collateral with correct bleed, resolution, and CMYK/RGB profiles.

Annotated 10,000+ text and image samples for NER, sentiment analysis, and intent classification for commercial NLP pipelines. Completed UHRS search quality evaluation and supported voice assistant training through transcription, speaker diarization, and emotion labeling, including English–Yoruba localization for multilingual model training.

Preprocessed and cleaned 50,000-row datasets with Python (Pandas, NumPy) for document classification and fraud detection. Assisted in fine-tuning a BERT-based text classifier, contributing to a 12% F1 score improvement, and built a Label Studio labeling pipeline that reduced annotation time by 35% while documenting evaluation findings and bias checks.

Grade: Second Class Upper (GPA: 3.6/5.0)

Activities and societies: Member of the Google Developer Student Club (GDSC); organized two AI workshops attended by 200+ students.

Earned a B.Sc. in Computer Science, with coursework spanning machine learning, natural language processing, software engineering, databases, and algorithms. Completed a final-year project building a text-based sentiment classifier for Pidgin English social media using fine-tuned DistilBERT.