Senior Solution Architect - Data, Ontology & AI Platforms

Overview

Requirements

• Own the end-to-end architecture of the data → semantic → AI stack and the contracts between layers (GraphQL, REST, protobuf, event streams, tool-calling interfaces)
• Lead and mentor the engineering team, reviewing designs and PRs against shared architecture decisions and reference patterns
• Architect the data platform — batch ingestion, a versioned lakehouse with branch-based safety (staging → quality gates → promotion), distributed transformation and orchestration, dataset versioning, and event-driven pipelines
• Design the semantic/ontology layer — entity, property, and relationship modeling; semantic query APIs; fine-grained policy enforcement; and dataset/column linkage that gives data meaning
• Drive the AI & MLOps layer — agent orchestration, AI copilot experiences, model guardrails, LLM inference, and the full experiment-tracking → model-registry → serving lifecycle
• Guarantee data and semantic quality — quality gates, ontology consistency, embedding/vector strategy, PII handling, and behavior evaluation that gates releases
• Embed governance by design — data lineage, auditability, multi-tenancy, and policy enforcement as first-class concerns
• Be the technical voice with stakeholders, translating data-governance and AI-operations requirements into architecture, and running design reviews and demos

Must-Have Qualifications

• 10+ years in software/data/ML engineering, with 5+ years as a solution, data, or platform architect leading teams and owning system-level design
• Deep distributed-data expertise — lakehouse architecture (e.g. Iceberg, Delta, Hudi), data versioning/branching, distributed compute (e.g. Spark), event-driven pipelines (e.g. Kafka), and data lineage/quality patterns
• Semantic / ontology / knowledge-modeling experience — entity-relationship or ontology design, semantic query layers, and GraphQL APIs; familiarity with policy-as-code for fine-grained access control
• MLOps and AI-agent systems — model registry and serving (e.g. MLflow, KServe), feature stores, LLM inference (e.g. vLLM), agent frameworks (e.g. LangGraph), and guardrails/safety patterns
• Vector search & retrieval — embeddings, semantic search (e.g. Qdrant, pgvector), and RAG patterns
• API and contract design — GraphQL, REST, protobuf, and event schemas with strong versioning discipline
• Hands-on coding in Python across the data/ML stack (e.g. FastAPI, PySpark)
• Strong technical leadership — mentoring engineers, writing architecture decision records, running design and PR reviews, and aligning a team around shared patterns
• Governance-by-design instinct — lineage, audit, multi-tenancy, and policy treated as first-class, not afterthoughts

Nice-to-Have

• Experience in regulated industries (financial services, healthcare, government) where data governance, lineage, and auditability are requirements
• Familiarity with tool-calling / agent integration protocols and contracts
• Background with declarative transformation frameworks (e.g. dbt) and data-quality tooling
• Working comfort with Kubernetes as a deployment substrate — enough to be effective, even if it is not your primary focus
• Familiarity with an observability stack (e.g. OpenTelemetry, Prometheus, Grafana) and immutable-audit patterns
• Prior experience taking a data/AI platform from proof-of-concept to production end-to-end

Benefits

• Data flows cleanly from source → versioned dataset → semantic layer → intelligent action, with quality and policy gates holding under real workloads
• The semantic layer becomes the trusted foundation that agents, copilots, and analysts query with confidence
• ML experiments, models, and agents ship predictably and safely, with lineage and audit intact
• Architecture boundaries and platform invariants hold as the system scales