Principal AI Architect

Principal AI Architect

Fulltime Position - US Remote (CST Time)

Job Description:

Principal AI Architect for client AssetMark

Generative AI, RAG and Enterprise Data/Application Architecture | Hands-on | 15+ years

Role Summary

We are looking for a senior, hands-on AI Architect to design, build and govern production-grade Generative AI solutions for enterprise clients. You will combine deep data and application architecture expertise with practical LLM engineering: RAG, vector search, agentic patterns and multi-model strategy across Anthropic, OpenAI and Gemini. You will own solutions end to end, from target-state architecture and ARB approval through secure, governed delivery on cloud.

Key Responsibilities

Define target-state AI, data and application architectures, reference patterns and reusable accelerators for enterprise GenAI programs.

Design and build RAG and agentic systems: ingestion, chunking, embeddings, hybrid retrieval, re-ranking, grounding, evaluation and observability.

Select and integrate LLMs (Anthropic Claude, OpenAI, Google Gemini) with sound model-routing, cost, latency, safety and fallback strategies.

Stay hands-on: write and review code, build proofs of concept, set engineering standards and unblock delivery teams.

Take architectures through Architecture Review Boards: prepare artifacts (ADRs, NFRs, logical and data architecture, integration flows), defend trade-offs and secure approvals on time.

Design security and access: RBAC/ABAC, identity integration, API management (APIM), secrets, network isolation, data classification, PII/PHI handling and audit trails.

Establish AI governance: prompt and model versioning, evaluation harnesses, guardrails, human-in-the-loop controls, lineage and responsible-AI practices.

Partner with data engineering on pipelines, semantic layers, metadata and data quality that make enterprise data AI-ready.

Advise client executives and engineering leaders, mentor architects and engineers, and contribute to pre-sales and solution shaping.

Required Qualifications

Experience: 15+ years in software, data and solution architecture, including at least 3 years designing and delivering production GenAI/LLM solutions.

Hands-on engineering: Strong Python (plus Java, TypeScript or similar), REST/GraphQL APIs and event-driven integration. Current in code, not only diagrams.

RAG and vector search: Demonstrated delivery of RAG systems with vector databases (e.g., Pinecone, pgvector, Azure AI Search, Vertex AI Vector Search, OpenSearch, Weaviate or Milvus). Retrieval quality tuning and evaluation.

LLM platforms: Hands-on with Anthropic, OpenAI and Gemini APIs, including tool/function calling, structured outputs, prompt engineering, context management and cost control.

Cloud: Production experience on GCP (strongly preferred), Azure or AWS: compute and containers (Kubernetes), managed AI services (Vertex AI, Azure OpenAI/AI Foundry, Bedrock), storage, serverless and CI/CD.

Data architecture: Data modelling, lakehouse/warehouse patterns, ingestion pipelines, metadata, lineage and data quality. Snowflake, BigQuery or Databricks experience.

Application architecture: Microservices, API design, integration patterns, resilience and non-functional requirements.

Security and governance: RBAC, identity (OAuth2/OIDC, Entra ID, IAM), APIM gateways, encryption and compliance (e.g., HIPAA, GDPR, SOC 2).

Architecture governance: Proven track record presenting to and securing approval from ARBs, with clear ADRs and trade-off analysis.

Communication: Clear written and verbal communication with executives and engineers. Experience leading without direct authority.

Preferred Qualifications

Agentic frameworks and orchestration (LangGraph, LangChain, LlamaIndex, MCP, Claude Agent SDK, Google ADK).

Knowledge graphs, ontology or GraphRAG (e.g., Neo4j).

LLMOps: evaluation frameworks, observability (e.g., LangSmith, Arize, Langfuse), guardrails and red-teaming.

Experience with AI-assisted engineering tools such as Claude Code for code generation and migration.

Cloud or architecture certifications (GCP Professional Cloud Architect, Azure Solutions Architect, AWS Solutions Architect, TOGAF).

Consulting or professional-services background, and experience building reusable accelerators.

Bachelor's or Master's degree in Computer Science, Engineering, Data Science or equivalent experience.