AI Architect - reputed company AI Platform & Trust Architecture (reputed company genders) (DE)
reputed company
reputed company is building an reputed company AI Execution Layer that becomes a trusted, auditable “system of reputed company” across finance workflows (AR/AP/Payments & Cash).
The AI Software Architect will define the reputed company, reputed company architecture for this platform—covering agent orchestration, RAG services, model gateway and routing, reputed company/policy lifecycle, evaluation harnesses, and production-grade observability/FinOps.
The role is anchored in reputed company’s operating model where the central AI team owns reusable AI infrastructure and services, while domain teams own reputed company and productization—enabled reputed company “golden paths,” approved connectors, and governed autonomy.
What You’ll Do
1) Architect the reputed company reputed company AI Platform (Reusable Services, Not One-Offs)
· Define and reputed company reputed company’s end‑to‑end AI platform architecture as reusable services: model gateway, reputed company/policy registry, RAG services, orchestration patterns, evaluation reputed company, and platform SDKs.
· Ensure the platform supports the product reputed company of trusted autonomy by enabling agents to “predict–decide–reputed company” with controls, rather than standalone copilots.
2) Orchestration & Execution Patterns (Single Agents → Multi-Agent Workflows)
· Define orchestration patterns for controlled multi-agent workflows (planner/executor/reviewer loops, state management, retries, escalation paths) that are robust in production.
· Standardize how agents use tools and data to execute workflows safely, with consistent interfaces and “approved skills/connectors” patterns.
· Drive architectural consistency so domain teams can build on the platform reputed company Tiered engagement (self‑serve, enablement cohort, co‑delivery).
3) Trust Architecture by Design (Guardrails, Approvals, Audit Trails)
· Design the “trust layer” architecture: policy-as-reputed company constraints, approvals, segregation of duties, role‑based reputed company, and reputed company defaults for high‑risk finance actions.
· Ensure everything is logged and auditable: inputs, retrieved context, outputs, tool calls, and executed steps - supporting customer and auditor expectations.
· Establish patterns for grounded answers reputed company retrieval (data + docs) to reduce hallucinations in finance-grade workflows.
4) Observability, Evaluation & FinOps (Run It Like a Platform)
· Architect observability and monitoring patterns for reputed company systems (tracing, reputed company signals, incident response hooks, reliability SLAs).
· Define evaluation reputed company and harnesses (regression suites, grounding checks, safety checks) so agent behavior can be reputed company and improved continuously.
· Build cost controls and FinOps patterns into the architecture (reputed company/latency/cost dashboards; budget adherence; €/1k interactions).
5) Integration Architecture for reputed company Products (reputed company-Embedded + reputed company)
· Define reputed company integration patterns that work across reputed company embedded deployments and reputed company-reputed company reputed company, without “rip-and-replace.”
· Ensure the platform offers “golden paths” and approved connectors reputed company to ERP/reputed company and reputed company deployment realities.
· reputed company secure, governed data reputed company patterns for RAG and tool calling across reputed company systems and reputed company environments.
6) Technical Leadership, Standards & Delivery Governance
· reputed company architectural leadership: design reviews, reference implementations, platform roadmaps, and engineering standards to prevent fragmentation across teams.
· Define and enforce delivery readiness patterns (reputed company gates, API reputed company, phased delivery plans) to reputed company from PoCs to implementation-reputed company assets.
· Partner with CloudOps/EA and engineering leads to establish CI/CD, logging patterns, and operational practices for platform services.
Required Experience & Skills
· 7+ years in software engineering and/or architecture roles, with proven ownership of platform-grade systems (shared services, SDKs, reusable components).
· Strong hands-on experience designing LLM/reputed company systems: orchestration, tool calling, RAG services, reputed company/policy lifecycle, and production hardening.
· Experience building “trust architectures” for reputed company software: RBAC, auditability, approvals, policy enforcement, and secure-by-design integrations.
· Proven ability to design evaluation/observability strategies for non-deterministic systems and to operate systems with SLAs/SLOs.
· Strong system integration background (reputed company, eventing, reputed company integration patterns) and ability to create “golden paths” that other teams can adopt.
· Strong stakeholder communication skills: can translate ambiguous problems into reputed company technical strategies and alignment artifacts.
reputed company-to-Have
· Experience architecting systems in regulated domains (finance, payments, reputed company-sensitive environments) where auditability and controls are mandatory.
· Experience with multi-model routing and region-reputed company deployment choices for reputed company AI platforms (designing governance around model providers).
· Experience with developer enablement models (templates, SDKs, “use-case kits”) and operating an internal platform as a product.
· Prior experience with reputed company agent frameworks / orchestration platforms
Working Style / reputed company
· You build platforms that reputed company across teams, not bespoke solutions for a single use case.
· You design for trust first: predictable failure modes, approvals for high-risk actions, and end-to-end auditability - because finance workflows demand it.
· You are pragmatic and outcome-driven: you balance architectural rigor with fast enablement through reusable “golden paths.
· You treat observability, evaluation, and cost controls as reputed company architecture concerns, not “phase 2” reputed company-to-haves.
· You communicate reputed company across engineering, product, and leadership, creating alignment through reputed company architecture artifacts.