Job Overview
We’re hiring an Enterprise Architect to define and govern the technical blueprint for an AI-first, data-centric platform. The role spans target architecture, reference patterns, and guardrails across orchestration, knowledge systems, retrieval pipelines, security, and operations. You’ll partner with product, security, data, and engineering leads to translate business outcomes into resilient, scalable designs-then guide teams to deliver them with measurable quality, cost, and reliability.
Key Responsibilities
- Target Architecture & Roadmap: Define the end-to-end architecture (experience, orchestration, agents/tools, data/knowledge, security, observability) and a phased roadmap.
- Reference Patterns & Standards: Establish patterns for multi-agent orchestration, connector/adapter design, model routing, policy-as-code, and evaluation/observability; produce reference implementations.
- Knowledge & Retrieval: Architect the coexistence of knowledge graph reasoning and retrieval (RAG) baselines; select graph/embedding stores, query patterns, and caching strategies.
- Integration & Platform Services: Define typed interfaces and contracts for AI engines, vector/graph stores, SQL/warehouse, and external tools; standardize retries, idempotency, and failure isolation.
- Security & Governance: Embed SSO/IDP, secret management, entitlement models (role/row/column), data minimization, auditability, and red-team controls into the design.
- Evaluation & SLOs: Institutionalize quality gates (faithfulness/correctness, precision/recall, latency), automated re-evaluation on change, and service SLOs with error budgets.
- Operability: Drive choices for tracing/logging/metrics, incident response, capacity planning, and cost/performance budgets; ensure CI/CD, IaC, and environment promotion gates.
- Technical Leadership: Review designs/PRs, mentor leads, resolve cross-team trade-offs, and communicate architecture decisions to technical and non-technical stakeholders.
Required Skills
- Architecture Mastery: 3-5+ years designing distributed systems or platforms (microservices, event-driven, API-first) on cloud environments.
- AI/LLM Platform Patterns: Practical experience with multi-agent orchestration, model/embedding selection, RAG pipelines, and knowledge graph solutions.
- Data & Knowledge Stores: Hands-on with one or more: graph DBs (RDF/SPARQL or property graph/Gremlin), vector DBs (pgvector, Milvus, Pinecone, Weaviate), SQL/warehouse.
- Security-by-Design: SSO/OAuth/OIDC, secrets management, least-privilege design, policy-as-code, audit logging, and data privacy controls.
- Interfaces & Contracts: Strong API design (REST/gRPC), typed schemas (OpenAPI/JSON Schema), error taxonomies, idempotency, and backoff strategies.
- Observability & Reliability: OpenTelemetry (traces/metrics/logs), call graphs, SLOs/error budgets, resiliency patterns (circuit breakers, bulkheads, retries, backpressure).
- DevOps & Delivery: IaC (Terraform/CloudFormation), CI/CD, artifact/version management, environment promotion and change control.
- Communication & Leadership: Clear decision records, stakeholder alignment, and the ability to mentor senior engineers and influence across teams.
Good to Have Skills
- Graph & Query Expertise: SPARQL, SHACL, Gremlin/PGQL; ontology/taxonomy stewardship and data lineage/provenance.
- Evaluation Tooling: Experience with LLM evaluation harnesses (e.g., custom scoring frameworks, judge models, A/B test rigs), regression dashboards, and automated re-eval pipelines.
- Model & Tooling Ecosystem: Familiarity with OpenAI, open-weights (Llama/Mistral), and connector protocols (e.g., MCP-style adapters).
- Runtime & Infra: Service mesh (Istio/Linkerd), API gateways, Kafka/event streaming, feature flags, multi-tenancy, and cost governance for model/runtime spend.
- Security & Compliance: Exposure to SOC 2/ISO 27001 (or comparable) practices; prompt-injection mitigation and data-exfiltration controls for LLMs.
- Languages & Stacks: Proficiency in one or more: Python, TypeScript/Node.js; comfortable reading/writing infra and application code to produce reference blueprints.
- UX & HITL Patterns: Designing clear approval flows and rationale/explanation UX for human-in-the-loop actions.
This role is ideal for a hands-on architect who can set a clear architectural runway, codify patterns, and steer multiple teams toward secure, explainable, and cost-effective AI capabilities at enterprise scale.