Object-relational mapping (ORM) frameworks simplify persistence, yet routine mapping choices—including inheritance strategy, association encoding, and denormalization—can substantially alter the generated SQL and the trade-offs among query latency, insert/update cost, and storage footprint. Existing optimization approaches either guarantee semantic validity but incur expensive per-candidate deployment and benchmarking, or learn from schema structure alone and therefore miss the fact that workload behavior is mapping-dependent. We present \textsc{TriORM}, a workload-aware neural–symbolic framework for multi-objective ORM mapping design that removes per-candidate workload execution from the \emph{online} recommendation loop while preserving validity by construction. \textsc{TriORM} (i) enumerates admissible mappings via bounded relational synthesis in Alloy, (ii) concretizes an abstract workload into schema-specific SQL templates for each candidate, and (iii) predicts continuous objectives using a tri-input model that fuses a typed schema-graph encoder, a concretized-workload encoder, and interpretable static cost proxies. The resulting predictions enable Pareto filtering and user-weighted selection without deploying or executing each candidate at inference time; profiling is performed only \emph{offline} to obtain supervision. On nine TradeMaker/\textsc{Leant} benchmark models, \textsc{TriORM} improves mean Pareto-front approximation over \textsc{Leant} (GD/HV $0.03/0.78$ vs.\ $0.07/0.61$) and reduces end-to-end recommendation time ($3.6{\times}10^3$s vs.\ $2.6{\times}10^4$s), while preserving semantic correctness within the chosen synthesis bounds.