Principal AI Engineer

We need to output HTML fragment with sections: About the role? The description includes Role Summary, Essential Functions, Responsibilities, Technical Capabilities, Technical Competencies, Behavioral Attributes, Location, Target Compensation, Schedule, Who We Are (benefits), Employment Statements (EEO etc). We must remove boilerplate: cookie/navigation/legal/EEO/equal-opportunity/scam-warning boilerplate, duplicated lines, application instructions. So we should drop EEO statements, "Who We Are" maybe includes benefits; we need to keep benefits if present. The "Who We Are" includes benefits list: Traditional medical, dental, vision coverage; Generous 401(k) match; Paid Time Off details; Paid federal holidays; Special employee pricing on lending products. That's benefits. So we should include a Benefits section with those bullet points. We need headings: About the role (maybe from Role Summary), Responsibilities, Requirements (maybe Technical Capabilities + Technical Competencies?), Qualifications (maybe Technical Competencies?), Skills (maybe Technical Capabilities?), Benefits, Pay, Schedule. The instruction: use

for section headings (e.g. About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule). So we need to decide which sections to include based on content. We have Role Summary (could be About the role). Essential Functions maybe part of Responsibilities? There's also separate Responsibilities list. We'll combine? Probably keep Responsibilities as separate section with bullet list. Technical Capabilities and Technical Competencies could be Skills and Requirements/Qualifications. Let's map: - About the role: Role Summary paragraph. - Responsibilities: the list under "Responsibilities" (the bullet-like sentences). Also maybe Essential Functions could be merged but we can keep as part of Responsibilities? The instruction: include a section ONLY when posting has real, specific content. We have both Essential Functions and Responsibilities; they overlap. To avoid duplication, we could merge them into one Responsibilities section, but we must preserve all substantive information. We could combine both lists into one

with items from both Essential Functions and Responsibilities, ensuring no duplication. Let's extract items. Essential Functions list (each bullet starts with a verb): - Own the technical direction of multiple automation and AI workstreams in partnership with AI Architect and other Senior Agentic AI Engineers - architecture decisions, tooling choices, and build quality. - Deliver complex automation and agentic AI systems from design through production deployment, including monitoring and post-deployment review. - Use RAG pipelines to connect agents to internal knowledge sources, ensuring accuracy and quality of agent responses. - Develop and enforce guardrails, safety layers, and human-in-the-loop verification patterns for agent outputs in a regulated financial services environment. - Build and maintain agent evaluation frameworks - moving beyond ad hoc testing toward systematic quality measurement, regression testing, and production observability for AI systems. - Mentor senior and mid-level engineers - code reviews, pairing, and active investment in their growth as engineers. - Partner with the Product Owner on discovery - helping scope automation opportunities technically, identify feasibility risks, and estimate effort honestly. - Develop and version structured prompt engineering standards - system prompts, output contracts, and verification rules - treating prompt design as an engineering discipline. - Contribute to knowledge sharing - rapidly bringing patterns and lessons learned into the broader IA engineering community. Responsibilities list: - Delivers automation and agentic AI systems independently, end-to-end, from discovery through production - Designs and maintains RAG pipelines, agent evaluation harnesses, and guardrails for production AI systems - Mentors senior and mid-level engineers - active, ongoing investment in their development - Partners with PO on discovery and scoping - technically credible contributor to the prioritization process - Drives AI-assisted development adoption - Develops and enforces prompt engineering standards and agent safety patterns We see overlap: many items similar. We'll combine unique items. Let's create a combined list: - Own the technical direction of multiple automation and AI workstreams in partnership with AI Architect and other Senior Agentic AI Engineers – architecture decisions, tooling choices, and build quality. - Deliver complex automation and agentic AI systems from design through production deployment, including monitoring and post-deployment review. - Use RAG pipelines to connect agents to internal knowledge sources, ensuring accuracy and quality of agent responses. - Develop and enforce guardrails, safety layers, and human-in-the-loop verification patterns for agent outputs in a regulated financial services environment. - Build and maintain agent evaluation frameworks – moving beyond ad hoc testing toward systematic quality measurement, regression testing, and production observability for AI systems. - Mentor senior and mid-level engineers – code reviews, pairing, and active investment in their growth as engineers. - Partner with the Product Owner on discovery – helping scope automation opportunities technically, identify feasibility risks, and estimate effort honestly. - Develop and version structured prompt engineering standards – system prompts, output contracts, and verification rules – treating prompt design as an engineering discipline. - Contribute to knowledge sharing – rapidly bringing patterns and lessons learned into the broader IA engineering community. - Delivers automation and agentic AI systems independently, end-to-end, from discovery through production (similar to deliver complex automation... maybe redundant) - Designs and maintains RAG pipelines, agent evaluation harnesses, and guardrails for production AI systems (similar to use RAG pipelines and guardrails) - Partners with PO on discovery and scoping – technically credible contributor to the prioritization process (similar to partner with PO) - Drives AI-assisted development adoption - Develops and enforces prompt engineering standards and agent safety patterns (similar to develop prompt engineering standards and guardrails) We need to avoid duplication but preserve all substantive info. Could keep both lists but risk duplication. The instruction: "Preserve ALL substantive information; do NOT invent or summarize away detail." So we must keep all info, but we can combine similar points if they are essentially same? Might be considered summarizing away detail if we merge. Safer to keep both sections: Essential Functions and Responsibilities as separate sections? But headings limited to given list; we could use "Responsibilities" for the Responsibilities list, and maybe "Essential Functions" not in allowed headings. So we could put Essential Functions content under Responsibilities as well, but that would be summarizing? Actually we would be adding more bullets to Responsibilities, preserving all info. That's okay: we just expand Responsibilities section with all bullets from both lists. That preserves all info, not summarizing away. So we will create a single Responsibilities section containing all bullets from Essential Functions and Responsibilities combined, deduplicating exact duplicates but keeping similar but not identical? We'll keep all distinct bullets; if some are very similar we could keep both to avoid losing nuance. Let's just include all bullets from both lists, even if some overlap; it's not summarizing away detail. Duplication is okay? The instruction says remove duplicated lines. So we should remove duplicated lines. So we need to detect exact duplicate lines. The lists have some overlapping meaning but not exact same wording. We'll keep both if not exact duplicates. Let's list all bullets exactly as written: Essential Functions: 1. Own the technical direction of multiple automation and AI workstreams in partnership with AI Architect and other Senior Agentic AI Engineers - architecture decisions, tooling choices, and build quality. 2. Deliver complex automation and agentic AI systems from design through production deployment, including monitoring and post-deployment review. 3. Use RAG pipelines to connect agents to internal knowledge sources, ensuring accuracy and quality of agent responses. 4. Develop and enforce guardrails, safety layers, and human-in-the-loop verification patterns for agent outputs in a regulated financial services environment. 5. Build and maintain agent evaluation frameworks - moving beyond ad hoc testing toward systematic quality measurement, regression testing, and production observability for AI systems. 6. Mentor senior and mid-level engineers - code reviews, pairing, and active investment in their growth as engineers. 7. Partner with the Product Owner on discovery - helping scope automation opportunities technically, identify feasibility risks, and estimate effort honestly. 8. Develop and version structured prompt engineering standards - system prompts, output contracts, and verification rules - treating prompt design as an engineering discipline. 9. Contribute to knowledge sharing - rapidly bringing patterns and lessons learned into the broader IA engineering community. Responsibilities: 1. Delivers automation and agentic AI systems independently, end-to-end, from discovery through production 2. Designs and maintains RAG pipelines, agent evaluation harnesses, and guardrails for production AI systems 3. Mentors senior and mid-level engineers - active, ongoing investment in their development 4. Partners with PO on discovery and scoping - technically credible contributor to the prioritization process 5. Drives AI-assisted development adoption 6. Develops and enforces prompt engineering standards and agent safety patterns Now check for exact duplicates: None are exact duplicates. Some similar but not same. So we keep all. Thus Responsibilities section will have 15 bullets. Now Technical Capabilities list: we need to decide where to put. Could be Skills section. Let's extract bullet points: Technical Capabilities: - Strong Python engineering skills - automation scripting, API integration, data processing, and production-quality code - Hands-on experience with LLM APIs (Claude, OpenAI, or equivalent) and agentic AI patterns - tool calling, orchestration, evaluation - Experience with agent orchestration frameworks (LangChain/LangGraph, CrewAI, n8n, or similar) - ability to evaluate tradeoffs and select appropriate tooling for the use case - Experience building RAG pipelines using vector databases (Pinecone, pgvector, Weaviate, Chroma, or similar) - Proficiency in structured prompt engineering - versioned prompts, system prompt design, output contracts, and evaluation-driven prompt iteration - Familiarity with Model Context Protocol (MCP) for agent-tool integrations, or demonstrated ability to learn tool-use standards quickly - Experience with cloud platforms (AWS, Azure, or GCP) for deploying and operating AI systems - Understanding of production requirements for AI systems - monitoring, drift detection, failure modes, and human-in-the-loop design - Comfortable with modern CI/CD practices and deployment pipelines - Familiarity with RPA tooling (UiPath or similar) - does not need to be a primary skill but must be able to contribute Technical Competencies: - 8+ years of software engineering experience, with at least 1–2 years working on automation, AI, or data engineering systems - Demonstrated experience delivering software to production - not just prototyping - Experience mentoring or providing technical leadership to less experienced engineers - Equivalent experience accepted in lieu of formal degree - Experience in a financial services or regulated environment a plus - Experience with agent observability tools (LangSmith, Arize, Langfuse, or similar) a plus - Familiarity with guardrails frameworks (NVIDIA NeMo Guardrails, Guardrails AI, or custom safety layers) a plus Behavioral Attributes: - Builds pod confidence through technical clarity - engineers in the pod always know what good looks like and how to get there - Promotes member experience as the measure of automation quality - Shows initiative in identifying automation opportunities beyond what the backlog already contains - Welcomes feedback and models the growth mindset expected of senior and mid-level engineers on the team - Open and energized by the pace of change in agentic AI - treats uncertainty as a feature, not a bug - Generous with knowledge - documents patterns, shares learnings, and contributes to the team’s collective intelligence Location: Marlborough/Chelmsford MA or Hillsborough, OR Target Compensation: $164,000 -$197,000 Schedule: 8am- 5pm Benefits from Who We Are: - Traditional medical, dental, and vision coverage - Generous 401(k) match - Paid Time Off: You'll accrue up to 15 days in your first year. In addition, you'll receive 40 hours of sick time and 3 personal days, which refresh annually - Paid federal holidays - Special employee pricing on lending products such as mortgage, auto, and personal loans (eligibility subject to standard account requirements and underwriting criteria) We need Benefits section with bullet list. Pay section