I'm Reynald Aryansyah, an AI Engineer focused on building reliable intelligent systems, from machine learning models to autonomous agent infrastructure.
Three perspectives on intelligent engineering: govern, orchestrate, understand.
Infrastructure for managing the AI agent lifecycle, with versioned configurations, evaluation, controlled execution, and traceable results.
Core enforces authority throughout the lifecycle. Bench precedes approval and publication.
Decision support for maritime logistics, combining cargo matching, voyage context, and document retrieval to produce recommendations that operators can review.
Branches depend on intent. Regulation uses retrieval; matching also uses graph and optimization. Transactions stay in Gateway.
Research on Indonesian meme classification using text-image fusion, topic information, and explainability to examine both predictions and model behavior.
Alternative experimental fusion paths. Topic features are used in A/B/C; variant D removes topic information.
I care about what happens around a model: how its inputs are prepared, how decisions are evaluated, and what happens when something fails. My work connects AI research with the engineering needed to make it understandable, testable, and maintainable.
About ↗