Senior AI Product & Systems Integration Architect

About the role

We are seeking a Senior AI Product & Systems Integration Architect to shape a confidential, stealth‑mode AI product. You will work hands‑on with a small technical team to translate vague AI concepts into concrete, buildable specifications and shipped features.

Key responsibilities

• Convert ambiguous AI feature ideas into detailed product and system specifications, user flows, data inputs, AI behavior, backend actions, and acceptance tests.
• Define agent capabilities, including read/write permissions, triggers, refusals, and exposure rules.
• Choose appropriate implementation approaches (LLM, retrieval, rule‑based code, automation, or human review) and design structured outputs, tool access, memory limits, fallbacks, escalation, and safety boundaries.
• Collaborate with UX, frontend, backend, data, AI, and QA teams to ensure end‑to‑end shippable features.
• Prototype using fixture data, mock agents, simulated events, and realistic state changes.
• Specify APIs, event contracts, queues, logs, permissions, schemas, and state transitions while protecting sensitive data.
• Design and execute tests for edge cases, hallucinations, privacy failures, refusals, and broken workflows.
• Freeze integration contracts before development and own the handoff through successful feature delivery.

Required profile

• Proven ability to work in ambiguous, fast‑paced environments and deliver complete, testable specifications.
• Experience turning high‑level AI concepts into production‑ready designs across product, UX, engineering, data, AI, and QA.
• Strong problem‑solving mindset with a focus on safety, privacy, and data protection.

Required skills

• LLM application architecture
• Agent workflow design and tool‑calling
• Structured output and schema validation
• Retrieval, context design, and memory boundary management
• API contract design and event‑driven system architecture
• Data modeling for AI‑driven features
• UX‑to‑system behavior mapping
• Fixture‑based prototyping
• Evaluation, regression testing, and acceptance criteria definition
• Privacy, permissions, audit, and data boundary enforcement