Job OverviewThe Principal Architect, Artificial Intelligence defines and drives architecture for our enterprise AI portfolio, including advisor-facing AI workspaces, a centralized AI control plane, and the agentic solutions built on top of both. This role provides architecture leadership across the end-to-end AI product development lifecycle (AI PDLC): discovery and intake, experimentation, design, build, evaluation, certification, deployment, operations, change management, and retirement.
You will establish repeatable governance and engineering patterns so that product teams can onboard to a common, governed path without reinventing controls for each initiative. The role pairs strategic architecture with hands-on depth: evaluating emerging models and frameworks, building reference implementations and proofs of concept, and guiding teams to deliver secure, resilient, observable, and cost-aware AI at scale in a regulated financial services environment.
Focus areas will be calibrated to the successful candidate's strengths and organizational priorities.
Key Responsibilities:AI Solution and Platform Architecture- Own and evolve reference architectures for AI-enabled applications, including user-facing AI workspaces and the agent runtime, data stores, and services behind them.
- Define patterns for durable user work: conversational and non-conversational sessions, memory and artifact continuity, graceful degradation, and integration into existing web platforms (identity, entitlements, feature flags).
- Architect intent routing, context assembly, and delegation between agents and domain services (A2A and task models), including human-in-the-loop controls and advisory-only boundaries so no unintended automated actions occur.
- Define experience architecture with product and UX partners: structured response rendering, latency budgets, streaming, error recovery, and multi-intent interactions.
- Partner on non-functional requirements: availability, deployment strategy, capacity, performance testing, and disaster recovery from pilot through scale-out.
Enterprise AI Control Plane and Integration- Define how AI products conform to a governed perimeter: API/SDK integration, AI gateway routing, policy enforcement, managed model inference, audit logging, and cost attribution.
- Architect registry and lifecycle patterns for models, prompts, tools (including MCP), and agents (e.g., Draft 1 Active 1 Suspended 1 Retired), including version promotion and kill-switch capability.
- Establish runtime controls: entitlements, governed tool and agent execution, rate limiting, pause/quarantine/kill mechanisms, and incident evidence capture.
- Specify reusable platform services, such as memory, knowledge and RAG abstractions, evaluation hooks, and observability contracts, that reduce duplication across teams.
AI Product Development Lifecycle and Governance- Lead or co-own the architecture of a repeatable path from use-case intake through production certification, ongoing change and exception management, periodic review, and retirement.
- Define classification and risk-tiering models that determine the evidence, evaluations, approvals, and identity steps required for each solution type.
- Establish production readiness gates: offline and runtime evaluation requirements, regression suites, red-team and safety standards, release evidence, and operational runbooks.
- Publish onboarding playbooks for internal teams and approved third parties, replacing one-off project checklists with consistent guidance.
- Align AI architecture with Enterprise Architecture, API and event review, security, risk, and compliance forums, and represent AI designs in architecture review cycles.
Security, Evaluation, and Operations- Define security architecture for AI solutions: SSO and token flows, scoped agent identity with delegated authorization, gateway validation, least-privilege tool access, PII protection at egress, content safety, and records retention for transcripts and generated content.
- Set standards for evaluation, including groundedness, hallucination, safety, and agent behavior, and for context and memory, semantic retrieval, and enterprise data access (authorization, lineage, abstraction).
- Lead LLMOps and AgentOps: versioning, deployment, monitoring, release management, FinOps (token and cost governance), and continuous improvement driven by production telemetry.
Technical Leadership- Define reference architectures for generative AI, agentic AI, RAG, tool-enabled and multi-agent solutions, and MCP/A2A interoperability.
- Run technology radar activities: POCs for models, agent frameworks, gateways, and evaluation tooling, and transition successful ones into reference implementations and platform roadmap input.
- Mentor solution and software architects, influence roadmaps, and communicate tradeoffs to senior product, engineering, and risk stakeholders.
What are we looking for?We're looking for strong collaborators who deliver exceptional client experiences and thrive in fast-paced, team-oriented environments. Our ideal candidates pursue greatness, act with integrity, and are driven to help our clients succeed. We value those who embrace creativity, continuous improvement, and contribute to a culture where we win together and create and share joy in our work.
Required Qualifications- 10+ years in enterprise, platform, or software architecture (distributed systems, cloud-native, data-intensive systems).
- 2+ years architecting generative AI, agentic AI, or ML platforms from experiment through regulated or enterprise production.
- 2+ years demonstrated ownership of AI product lifecycle outcomes, spanning governance, security, evaluation, and operations in addition to models.
- 2+ years hands-on credibility with AI applications, agents, RAG, APIs, integrations, and technical POCs, and the ability to review code and designs with engineering teams.
- 7+ years strong command of cloud-native design: microservices, APIs, event-driven patterns, containers/Kubernetes, IAM, observability, resilience, and performance.
Preferred Qualifications- Depth in AWS: EKS, Amazon Bedrock, API Gateway, IAM, data stores (e.g., PostgreSQL, DynamoDB, S3), and observability tooling.
- Experience with AI gateways and OpenAI-compatible inference APIs.
- Familiarity with agent protocols and frameworks: tool calling, MCP, A2A, multi-agent orchestration, embeddings, and vector search.
- Experience in financial services or other highly regulated domains: supervisory records, advice-related content controls, entitlement-heavy systems.
- Knowledge of enterprise identity platforms, non-human identity (NHI) for agents, and delegated authorization patterns.
- Proficiency in Python, Java, IaC, CI/CD, and DevSecOps; strong architecture documentation skills (C4, sequence and data-flow diagrams, ADRs).
- Bachelor's or Master's degree in Computer Science, Engineering, AI, or a related field.
Core Competencies
- AI solution and user-experience architecture
- Enterprise AI platform and control-plane integration
- AI PDLC, certification, and lifecycle governance
- Generative and agentic systems; RAG and enterprise knowledge access
- AI security, Responsible AI, and runtime policy enforcement
- Evaluation, quality engineering, and continuous assurance
- LLMOps/AgentOps, observability, resilience, and AI FinOps
- Architecture governance, standards, and stakeholder leadership
Pay Range: $155,942.00 - $259,869.00
Actual base salary varies based on factors, including but not limited to, relevant skill, prior experience, education, base salary of internal peers, demonstrated performance, and geographic location. Additionally, LPL Total Rewards package is highly competitive, designed to support your success at work, at home, and at play - such as 401K matching, health benefits, employee stock options, paid time off, volunteer time off, and more. Your recruiter will be happy to discuss all that LPL has to offer!
