AI
AI systems, not just AI demos.
Bernard is interested in systems where AI interacts with tools, APIs, workflows, documents and infrastructure — not isolated chat interactions.
AI Engineering
Designing systems where models can use tools, call APIs and operate with deliberate constraints.
Agentic Systems
Connecting models to workflows, data sources and external systems without fragile one-off prototypes.
RAG
Grounding model responses in structured and unstructured information as part of a larger system.
Tool Integration
APIs, databases, files and document workflows are part of the product surface, not afterthoughts.
AI Infrastructure
Observability, evaluation and operational reliability matter as much as model capability.
AI Security
Data boundaries, credentials, permissions and safe operational assumptions are essential.
System view
User
↓
Agent
↓
Model
↓
ToolsAPIsDatabaseFilesWeb
↓
Verification / Observability
↓
Result