AI Engineer

• Design clear prompt templates and role scaffolds (System/User/Assistant).
• Keep chats relevant with windowing/summarization and retrieval-backed long-term memory; basic tenant/privacy scoping.
• Chunk/clean docs, embed into a vector store, run hybrid retrieval (BM25 + dense), and return answers with citations.
• Favor evidence, cite sources, and abstain when unsure.
• Implement Plan → Act → Observe → Iterate; call APIs/DBs via typed tool contracts.

Applied AI Engineer

Semantics Architect – Data & AI Platform