We build AI systems that deeply understand specialty insurance - actuarial concepts, underwriting logic and pricing frameworks embedded into production products like automated triage, intelligent document extraction and pricing recommendations that actuaries actually trust.

We own the infrastructure that makes AI reliable at scale - from evaluation frameworks that measure model performance across thousands of insurance scenarios, to deployment pipelines that safely roll out new models to customers. We balance cutting-edge research with production constraints like latency, cost and compliance.

We design and deploy multi-agent systems that coordinate across tools, memory and knowledge bases. From routing complex insurance queries to orchestrating document processing pipelines, we build agents that make autonomous decisions while maintaining reliability and explainability in high-stakes environments.

AI at hx builds production models from scratch when off-the-shelf solutions don't meet requirements. We train custom models on proprietary insurance data, experiment with novel architectures, and push beyond what's available through APIs - treating model development as core engineering work, not just prompt tuning.