Rakib Hassan

I developed and evaluated research methods in conformal prediction, uncertainty quantification, and temporal knowledge graph forecasting through independent machine learning research. My work includes a history-aware ranking adapter that uses candidate recurrence, recency, frequency, and staleness, evaluated across forecasting architectures and benchmark datasets.

I also developed Aegis, an ML-based fraud risk monitoring and decision support platform, integrating calibrated ensemble inference with REST APIs and PostgreSQL-backed workflows. I have coauthored three conference papers and am interested in LLM engineering, retrieval-augmented generation, knowledge-grounded reasoning, and reliable AI systems.