Afareayo Soremekun

I’m a Data Analyst who builds end-to-end data systems—turning messy inputs into reliable datasets, insights, and customer-facing outcomes. I focus on measurable performance gains: faster queries, higher data quality, lower costs, and clearer decision-making.

At Parley (Stealth), I optimized SQL performance using indexing and stored procedures, then led product development for an AI-powered platform with conversational search, automated price negotiation, and real-time booking. I also built an end-to-end hotel data extraction pipeline in Python (processing 35,000+ sites) and achieved 86% data completeness using OpenAI GPT-4 with intelligent chunking.

I’ve strengthened real-world data reliability with practical engineering: I implemented Hugging Face SigLIP-based image classification and filtering across 125,000+ images, including perceptual duplicate detection and relevance scoring to reduce storage by 37% without sacrificing relevant images. I designed monitoring and evaluation with real-time progress tracking, detailed error reporting, and web-based dashboards, and improved private beta user satisfaction by 22% after feedback-driven optimization.

Before that, I delivered business impact across CRM operations, automation, and analytics—migrating data into cloud databases (cutting compilation/querying time by 90%), building automated workflows that transformed LinkedIn data into Google Slides reports (10x faster), and deploying a containerized Streamlit document automation app that reduced contract generation time from 20 minutes to seconds while eliminating human error. In my earlier roles, I produced Power BI reporting at scale, accelerated insight delivery by 95%, and used Python + scikit-learn (including KMeans) to recover value—such as a customer churn model that reduced churn by 17% and segmentation that contributed to a 26% increase in tier 1 customers.