| Job Summary | - The Lead Applied AI Engineer designs, builds, deploys, and continuously improves secure, scalable, and supportable AI solutions on MUFG AIRP. The role applies foundation models, LLMs, retrieval-augmented generation (RAG), and agentic AI patterns to high-value business problems in a regulated financial-services environment.
- This is a senior hands-on technical leadership role. The incumbent turns AI concepts and prototypes into production capabilities, applies reusable engineering patterns, and partners across product, architecture, cloud/platform, security, risk, compliance, data, and operations.
|
| Key Responsibilities | - Lead the design, build, deployment, and continuous improvement of production LLM, RAG, and agentic AI solutions for priority MUFG business workflows.
- Translate business needs into secure, scalable, and supportable AI solution designs with clear success measures, use boundaries, human-oversight requirements, and operational ownership.
- Design and optimize application-specific RAG implementations using approved AIRP capabilities, including document onboarding, metadata, chunking, retrieval configuration, reranking, grounding, citations, and source traceability.
- Design controlled, business-specific agentic workflows using approved orchestration and secure tool-execution patterns, including validation, least-privilege access, error handling, and human approval where required.
- Select and apply appropriate models, prompts, context strategies, retrieval approaches, and integration patterns based on quality, risk, latency, resilience, cost, and business requirements.
- Define and execute evaluation and release-readiness practices for task quality, groundedness, retrieval relevance, safety, reliability, latency, and cost.
- Build and operate application services, APIs, and enterprise integrations using AIRP-approved model access, cloud, container, CI/CD, security, and observability capabilities.
- Improve application performance through prompt and context optimization, model routing, caching, batching, model adaptation, fine-tuning, quantization, or inference optimization where appropriate.
- Own the application-level AI lifecycle for assigned solutions, including versioning, evaluation gates, controlled releases, monitoring, incident response, rollback, and continuous improvement.
- Embed security, privacy, Responsible AI, model-risk, data-protection, and audit requirements into each solution, partnering with relevant control functions as needed.
- Apply AIRP reference architectures and engineering standards; contribute reusable application patterns, implementation feedback, and platform enhancement requirements to the AIRP roadmap.
- Provide technical leadership for assigned initiatives through design reviews, mentoring, documentation, and collaboration with product, architecture, platform, cybersecurity, data, risk, compliance, and operations teams.
|
| Required Qualifications | - Experience in banking, financial services, risk, compliance, financial crime, operations, or another regulated enterprise environment.
- Experience with AWS Bedrock, SageMaker, OpenSearch, Azure OpenAI, Vertex AI, Databricks, vLLM, Triton, MLflow, Kubeflow, model gateways, or private/open-source LLM deployments.
- Experience with KYC, underwriting, financial crime, sanctions screening, document intelligence, knowledge retrieval, financial analysis, or workflow automation.
- Familiarity with AI governance, Responsible AI, model risk, privacy, audit controls, vendor risk, and AI cost governance.
- Experience applying evaluation frameworks, red-team/adversarial testing, release gates, and post-production AI monitoring.
|