Selected work / 2022–2026
Things I've built.
Research, experiments, and working applications. Each project starts with a problem worth understanding.
01Applied AI / EvaluationSingle-agent vs. multi-agent LLM systemsDesigning, orchestrating, and evaluating agent architectures for financial numerical reasoning.
02Machine learning / FastAPIInsurance Charges PredictorAn end-to-end regression system with a live API-backed web application.
03Explainable AI / AnalyticsCustomer Churn Prediction with SHAPA churn-classification workflow with transparent predictions and feature attributions.
04Business analytics / MLBlinkit Sales Performance PredictionUnderstanding the relationship between marketing inputs and sales through regression.
✳
Additional experiments include RAG document Q&A and applied deep learning. More case studies will be published as they are ready.