[Remote] AI Forward Engineer

Note: The job is a remote job and is open to candidates in USA. Nuuvia is the leading provider of intelligent lifecycle banking solutions for community banks and credit unions. They are seeking an AI Forward Engineer to embed AI into their operations and work closely with credit union partners to develop AI-powered solutions that enhance customer engagement and compliance within the financial sector.

Responsibilities

• Embed with credit union partners to identify high-value AI use cases — youth onboarding, financial coaching, support deflection, fraud signal detection, internal ops automation
• Translate field requirements into production AI features that ship to that FI within weeks, not quarters
• Customize AI workflows per institution while preserving a reusable core balance with generality
• Run AI-enablement sessions with FI leadership, ops, and member service teams
• Operate as the technical face of Nuuvia AI to credit unions: requirements gathering, demos, joint design reviews, and post-launch iteration
• Own model selection: match the right model to each task based on capability, cost, latency, privacy posture, and regulatory risk
• Build agent and RAG pipelines using frameworks such as LangChain, LlamaIndex, Semantic Kernel, Microsoft Promptflow, or in-house equivalents
• Implement prompt engineering, function calling, tool use, and multi-step agent patterns hardened for production reliability
• Maintain a model registry — track which models are in use, for what purpose, which version, and last evaluation date
• Monitor for model drift, hallucination rates, and output degradation; drive measurable cost efficiency through token budgeting, caching, batching, prompt compression, and smart model routing
• Design and implement guardrails for every AI-assisted workflow: prompt injection prevention, PII detection and masking, output filtering, and content safety
• Build audit trails and logging for all AI interactions — every prompt, every response, every action taken — to support regulatory examination
• Implement human-in-the-loop controls for AI-assisted decisions with regulatory exposure (member-facing content, account actions, eligibility logic)
• Apply industry-standard LLM security frameworks as a baseline across all AI tooling
• Ensure no member PII flows through external model APIs without explicit anonymization or approval
• Operate with full awareness that Nuuvia serves federally regulated financial institutions — compliance is a design constraint, not a blocker
• Partner with internal compliance and external regulators to ensure AI-generated content and AI-assisted decisions meet documentation, explainability, and audit requirements
• Ensure AI-generated content reaching members or affecting account decisions can be explained in plain language
• Document AI model behavior, known limitations, and risk mitigations to a standard appropriate for regulated examination
• Treat audit readiness as a continuous practice — automated evidence collection, control testing, and policy enforcement around AI systems
• Extend NautBot — Nuuvia's internal AI agent that automates engineering ops, monitoring sweeps, ticket triage, and routine workflows across Microsoft Teams, Jira, Datadog, and Azure — and expand its coverage to additional client-facing surfaces
• Build reliable, observable automation pipelines with full traceability
• Drive Chat Action Center automation for day-to-day workflows and tasks
• Automate internal and client-facing workflows — Jira blocked-ticket detection, sprint automation, story-point estimation, escalation routing, deployment alerts, incident summaries
• Ensure all automated client-facing messages are accurate, auditable, and contextually appropriate
• Treat every deployed AI feature as an evolving system: instrument feedback, watch real usage, and rapidly iterate
• Translate field signal into product priorities: what worked, what failed, what regulators flagged, what FIs asked for
• Partner with Engineering, Product, Implementation, and CSM teams to keep client-specific work from forking the core platform

Skills

• Bachelor's degree in computer science, Engineering, or related field (Master's a plus)
• 4–7 years of software engineering experience, with at least 2 years shipping AI/LLM-powered systems to production
• Strong Python (primary) and/or TypeScript. Comfortable across the full stack when needed
• Hands-on experience with LLM APIs — prompt engineering, function calling, tool use, agent patterns
• Demonstrated experience building guardrails or safety systems for AI: PII masking, output filtering, audit logging
• Practical understanding of model risk management — documentation, validation, monitoring
• Experience with at least one relevant framework: LangChain, LlamaIndex, Semantic Kernel, Guardrails AI, or Microsoft Promptflow
• Solid security fundamentals: OAuth 2.0, secret management, least-privilege API access
• Experience with Azure (App Services, Azure OpenAI, Key Vault, Monitor) or equivalent cloud platform
• Comfort working in a regulated industry — or proven ability to learn fast and partner with compliance teams to ensure AI systems meet regulatory expectations
• Comfort working in a small, senior team with minimal layers — no project managers, no ticket groomers, no handholding
• Customer-facing maturity: can run a meeting with a credit union CIO, COO, or fraud officer and walk out with aligned next steps
• Forward Deployed Engineer (FDE) background — Palantir, Sierra, Glean, Anthropic Solutions, OpenAI Forward Deployed, Brex Forward Deployed, or equivalent customer-embedded AI engineering role
• Experience with Azure OpenAI Service and Azure AI content filtering/safety features
• Familiarity with model evaluation frameworks (LangSmith, PromptFlow Evals, custom eval pipelines)
• Experience with PII detection and masking tools (Microsoft Presidio, AWS Comprehend, or similar)
• Prior experience in a regulated industry (financial services, healthcare, government)
• RAG patterns + vector search (Azure AI Search, Pinecone, pgvector)
• Microsoft Teams bot / connector development
• Experience with general-purpose agent harnesses (OpenClaw, Hermes, or equivalent)
• Fine-tuning experience (LoRA/PEFT) — nice to have, not required

Company Overview