Practice

Agentic AI engineering

AI shopping assistants, customer-support agents and the infrastructure behind them: orchestration, memory, resilience and observability.

Scope

Applied capabilities

Python services connected to ecommerce APIs for product advice and customer support, with operator review and escalation in the support workflows
Capability-based multi-agent routing and tool-enabled workflows
Circuit breakers, dead-letter queues, offline buffers, and recovery
Local LLM routing, GGUF operations, and resource-aware model loading
Graph and vector memory, event buses, and real-time interaction
Pilot design, controls, evidence, and governance for AI adoption

In practice