Oladapo Abimbola's Work | ContraWork by Oladapo AbimbolaOladapo Abimbola
Data Analyst | Excel | SQL | Power BI | Python | AI Builder
Project Title:
Repurpose — One Idea, Three Platform-Ready Posts
Project Link:
https://app.flowstep.ai/prototype?activeFileId=765b65ce-0b6d-4d85-b86b-b33dd2c83ce9
Project Description:
Repurpose is a mobile-first content repurposing tool built for solo, faceless content creators. Most creators lose hours turning one video script into platform-specific posts, because every platform needs its own tone and format. Repurpose solves that: paste your script or idea, tap Generate Variants, and get a TikTok hook, a YouTube Shorts caption, and a Facebook Reel description — each written for that platform's audience. One idea in, three platform-ready posts out. Designed with a clean, minimal, Linear-inspired interface (near-black background, single violet accent, no clutter) so it feels like a real daily-use app rather than a mockup. Built entirely in Flowstep.
Walkthrough Video:
Completed coursework for DataTalksClub's LLM Zoomcamp 2026, covering the core building blocks of production RAG/LLM systems: Agentic RAG (converting OpenAI-based notebooks to Groq API), Vector Search with sentence-transformers and minsearch, Orchestration with Kestra (workflow automation in a containerized environment), and Retrieval/LLM Evaluation using Hit Rate and MRR metrics across text, vector, and hybrid search approaches. Debugged real-world integration issues including API compatibility, deprecated model migrations, and credential rotation.
GitHub :https://github.com/BimbsyStats/llm-zoomcamp-2026
https://github.com/BimbsyStats/llm-zoomcamp-hw5 Completed Excelerate's Data Visualization Associate internship (Team 29) — a 4-week program building a compliance monitoring pipeline for Saint Louis University's international student data. Worked across three datasets (Applicant: 6,894 rows; Connect: 34,341 rows; SEVIS: 3,524 rows), setting up a PostgreSQL/pgAdmin → Supabase → Looker Studio pipeline. Conducted SQL-based exploratory analysis, built a 12-chart compliance dashboard tracking metrics like financial compliance (45.46%) and I-20 issuance rate (34.68%), and caught a critical data quality issue affecting 624 records. Presented as Speaker 3 (Dashboard & Validation) in the final team presentation. Completed with a 93% overall evaluation and awarded "Star Performer" certificate. OrderBae AI
OrderBae is an AI-powered web application built to help fashion designers and small boutiques manage customer inquiries, showcase products, and simplify the ordering process.
The platform provides an organized product catalog, an intelligent AI assistant for answering customer questions, and a streamlined shopping experience. It helps reduce repetitive conversations on WhatsApp and social media while making it easier for customers to discover products and place orders.
Key features
AI-powered customer assistant Product catalog and category management Smart product search Order inquiry support Mobile-friendly web application Designed for fashion businesses and boutiques
Technologies
Python AI/LLMs Web Application Development Git & GitHub
This project demonstrates my ability to combine AI with practical business workflows to create tools that improve customer experience and support business growth.
GitHub :https://github.com/BimbsyStats/OrderBae---Al-Fashion-Assistance