Semantic Search & RAG
Designed retrieval workflows around embeddings, vector search and LLM generation to improve grounded responses.
I’m David Taiwo Balogun, a Senior Data Scientist and AI-focused data professional building analytics, machine learning and intelligent data products.

David Taiwo Balogun is a Data Scientist, AI Researcher, AI Expert, and Lead Analyst with experience developing machine learning solutions, credit risk models, and real-time analytics platforms across fintech, financial services, edtech, and technology-focused organizations.
He specializes in the intersection of artificial intelligence, financial risk, data analytics, and data engineering, with experience leading end-to-end projects spanning data engineering, predictive modelling, business intelligence, dashboard development, and stakeholder engagement. His work focuses on transforming complex data into actionable insights that improve decision-making, operational efficiency, and risk management.
David is also an active AI researcher and has authored and presented research at international conferences, including the 10th International Conference on Cloud and Big Data Computing (ICCBDC 2026) at the University of Cambridge, the 10th International Conference on Advances in Artificial Intelligence (ICAAI 2026) in London, and the International Conference of The Faculty of Computing (ICFC 2026), covering areas such as low-bandwidth AI systems, computer vision for agriculture, and Retrieval-Augmented Generation (RAG).
Beyond industry and research, David is passionate about developing the next generation of data professionals. He has trained and mentored nearly 3,000 aspiring data scientists and analysts across Africa, Europe, and the United States through technical education, mentorship, and community initiatives.
By combining technical expertise, research, and business acumen, David uses data science and AI to solve complex problems, advance applied AI research, and deliver measurable business impact.
A snapshot of the scale and outcomes behind my work across data, AI, analytics and education.
A mix of machine learning, analytics engineering, AI systems and applied research.
Designed retrieval workflows around embeddings, vector search and LLM generation to improve grounded responses.
Built predictive modelling workflows for financial risk, classification and decision support in lending environments.
Applied deep learning to image classification, from dataset preparation and training through evaluation.
Built dashboards and reporting systems that turn loan and asset-financing data into faster stakeholder decisions.
Research on designing an end-to-end pipeline from data collection to deployment in constrained environments.
Explored practical approaches to responsible and transparent machine learning through open-source work.
I care about the entire system—not just the model. Data quality, retrieval, evaluation, deployment and measurable outcomes all matter.
Leading data analytics initiatives across lending and asset financing. Built dashboards and decision-support systems that improved real-time stakeholder visibility and supported £24.8M+ in loan disbursements.
Worked across financial data, reporting and analytics to support product and business decisions.
Applied data science to lending and portfolio problems, contributing to a 10% reduction in non-performing loans.
Designed and delivered technical training, bootcamps, mentoring and talks covering data analytics, Python, SQL, machine learning and AI.
Peer-reviewed and conference research spanning Edge AI, retrieval-augmented generation, computer vision for agriculture and deep learning for financial risk.
Presented virtually at the 10th International Conference on Cloud and Big Data Computing (ICCBDC 2026). The work proposes a modular device–edge–cloud pipeline for resource-constrained environments.
A multi-crop dataset of 14,850 PlantVillage images across 4 crops and 9 disease classes was used to benchmark a scratch CNN against MobileNetV2, ResNet-50 and EfficientNetB0 transfer-learning models for mobile deployment in West Africa.
Full paper accepted for presentation and publication at the 10th International Conference on Advances in Artificial Intelligence (ICAAI 2026), London. The work studies how chunking and retrieval design choices reduce hallucination in domain-specific RAG systems.
Presented at the International Conference on Intelligent Technology, System and Service for Internet of Everything (ITSS-IoE 2026), Babcock University. The work explores a lightweight CNN–Transformer hybrid architecture for real-time acoustic event detection on edge gateways, targeting Industrial Internet of Everything deployment.
Paper publication in progressPresented at the 5th International Conference on Artificial Intelligence and Robotics, University of Lagos. The work audits dataset bias across AI systems in African contexts, examining fairness and representational gaps in real-world AI deployments.
Paper publication in progressExamines LSTMs, GRUs and Transformers for credit default prediction, alongside ensemble approaches combining neural and classical models for robust, explainable risk assessment.
Read full paper ↗Grouped by where the work was published.
Selected publications, television appearances and thought-leadership features covering data, AI and technology.
Featured television appearance discussing technology, data and AI.
Watch on YouTube ↗Featured on TVC in a technology-focused discussion.
Watch on YouTube ↗I'm interested in thoughtful products, applied AI and difficult data problems.