Remote Data Science Lead Jobs

Applied AI at First Citizens is not about fitting every problem to the nearest large language model, nor about chasing the latest foundation model release. It is about understanding a meaningful business problem well enough to know what kind of solution it actually calls for — which might be a predictive model, a decision tree, a classical statistical method, a natural language processing approach, a generative AI solution, or no AI at all. That last option is not a failure; it is good judgment.

Not every problem needs AI. Not every AI problem needs GenAI. Holding that standard — and being willing to say so — is central to how this team earns credibility and delivers durable value. Success here is measured by fit-for-purpose solutions, disciplined evaluation, responsible implementation, and outcomes that the bank can actually sustain and stand behind.

The Applied AI / Data Science Lead will provide hands-on execution capacity across data science and generative AI engineering. The role works closely with business product owners, data and technology teams, AI platform partners, and Responsible AI and risk stakeholders to shape use cases, build solutions, establish evaluation methods, and support the path from experimentation to production. This is a senior professional individual contributor role — someone who can independently lead complex technical work, make sound modeling and design decisions, and communicate trade-offs clearly to stakeholders at multiple levels.

Business Problem Framing and Use Case Shaping

• Work with business leaders and product owners to identify, assess, and shape high-value data science and AI opportunities across the General Bank, Commercial Bank, and Enterprise Functions.
• Translate business questions into well-defined analytical problem statements, with clear success measures, data requirements, solution hypotheses, implementation considerations, and expected value outcomes.
• Assess whether a problem is best addressed through conventional analytics, statistical modeling, machine learning, generative AI, workflow change, or no AI solution at all — and recommend a fit-for-purpose approach grounded in evidence and practicality, not novelty.
• Support prioritization of AI use cases by evaluating business value, data readiness, implementation feasibility, risk and control implications, operating model requirements, and the ability to measure impact over time.

Solution Design and Hands-On Development

• Design, build, validate, and refine analytical and AI solutions using appropriate methods: predictive modeling, supervised and unsupervised machine learning, natural language processing, generative AI, retrieval-augmented generation, optimization, or other advanced analytics techniques — selected on the basis of fit, not fashion.
• Develop data pipelines, features, model prototypes, prompt or retrieval configurations, evaluation datasets, reusable code assets, and supporting documentation required for experimentation and responsible implementation.
• Establish transparent baselines and, where warranted, challenger approaches so that solution complexity is justified by measurable performance improvement or business value — not technical preference alone.
• Contribute technical judgment on model selection, vendor capabilities, enterprise platform services, solution architecture, integration needs, and production-readiness considerations.

Evaluation, Measurement, and Responsible Delivery

• Define and execute fit-for-purpose evaluation plans covering model performance, stability, interpretability, robustness, data quality, user acceptance, operational feasibility, monitoring, and business outcome measurement as appropriate to each use case.
• For generative AI solutions, develop evaluation approaches for task accuracy, groundedness and faithfulness, retrieval quality, human review effectiveness, harmful output risk, prompt handling, and other use-case-specific performance and control requirements.
• Partner with Responsible AI, model risk, business risk, compliance, legal, cybersecurity, privacy, and other stakeholders to ensure solutions are developed with appropriate documentation, testing evidence, controls, and ongoing monitoring plans from the start — not retrofitted at the end.
• Clearly communicate model assumptions, limitations, trade-offs, risks, recommended controls, and decision implications to business and technical stakeholders in language that is accessible, not just technically accurate.

Enterprise AI Platform and Reusable Capabilities

• Work alongside AI platform and technology partners as the bank's enterprise AI capabilities mature — providing practical requirements from data science delivery and positioning solutions to leverage approved platform services when ready.
• Develop reusable design patterns, evaluation methods, templates, code assets, and delivery best practices that help the bank build AI solutions more consistently, securely, and efficiently over time.
• Support technical evaluation of AI tools, technologies, and vendors through objective testing and structured assessment of their relevance to business needs, enterprise architecture, and responsible adoption requirements.
• Contribute to experimentation and implementation pathways that connect enterprise data, AI models, monitoring, governance evidence, and operational workflows — building the infrastructure for AI at scale, not just one-off solutions.

Stakeholder Partnership and Technical Leadership

• Collaborate across business, data, technology, architecture, platform, and risk teams to move use cases from early ideas to disciplined experimentation and appropriate implementation — navigating complexity without losing momentum.
• Present analytical findings, solution alternatives, technical recommendations, risks, and outcomes in clear language for senior partners and decision makers who may not have a technical background.
• Share knowledge, mentor less experienced analysts through project delivery, and contribute to a team culture built on curiosity, craft, rigor, and honest evaluation of what is working and what is not.
• Stay current with meaningful developments in AI, ML, GenAI, and advanced analytics while maintaining a pragmatic focus: understanding what is actually ready for enterprise adoption versus what is still better suited to a research paper.

WHAT SUCCESS LOOKS LIKE

• High-value business problems are translated into sound analytical or AI use cases with clear success measures and realistic implementation paths — including a clear view of what "good enough" looks like and when more complexity is not warranted.
• Solutions use the right technique for the problem. Model complexity and generative AI are justified by evidence of better outcomes, not by the availability of new technology.
• Experiments are designed rigorously, documented clearly, and positioned for responsible implementation through genuine partnership with platform, technology, and risk teams.
• Reusable analytical patterns, evaluation methods, and delivery assets help First Citizens increase speed, consistency, and trust as the bank builds its AI capability over time.

PREFERRED QUALIFICATIONS AND SKILLS

• Experience developing, evaluating, and productionizing data science, machine learning, advanced analytics, and/or generative AI solutions that address real business problems in financial services or another regulated enterprise environment.
• Strong applied knowledge of statistical modeling, supervised and unsupervised machine learning, natural language processing, experimental design, model evaluation, performance monitoring, and production model lifecycle practices.
• Hands-on experience with Python, SQL, and common data science / ML libraries and frameworks, with the ability to build reusable, maintainable code assets for experimentation, evaluation, deployment support, and monitoring.
• Practical experience building or supporting AI/ML solutions on cloud and enterprise data platforms, including AWS and Snowflake. Experience with AWS Bedrock for foundation-model application development and Amazon SageMaker for custom ML model development, deployment, monitoring, and lifecycle management is strongly preferred.
• Experience designing or implementing generative AI solutions using retrieval-augmented generation, vector databases, GraphRAG or knowledge-graph-enhanced retrieval, model orchestration, prompt and context management, and structured output generation.
• Experience with Snowflake Cortex AI, Snowflake-native AI capabilities, or comparable enterprise AI services is strongly preferred, especially in use cases involving document analysis, text classification, summarization, semantic search, or business workflow enablement.
• Experience with agentic AI or tool-using LLM patterns, including multi-step workflows, conditional routing, tool/function calling, human-in-the-loop escalation, and observability. Familiarity with frameworks such as LangGraph, LangChain, LlamaIndex, or comparable orchestration tools is valuable.
• Experience defining and executing evaluation strategies for GenAI and ML solutions, including task accuracy, retrieval quality, groundedness, faithfulness, hallucination risk, robustness, latency, cost, safety, human review effectiveness, and post-production monitoring.
• Experience with monitoring, observability, and production support practices for AI/ML or GenAI systems, including model performance monitoring, drift detection, logging, tracing, alerting, error analysis, and operational feedback loops. Experience with tools such as SageMaker Model Monitor, Clarify, Langfuse, or comparable monitoring / observability platforms is valuable.
• Ability to partner effectively with platform engineering, enterprise architecture, data management, cybersecurity, model risk, Responsible AI, technology risk, and business teams to deliver technically sound, scalable, secure, and well-controlled AI solutions.
• Strong communication, documentation, stakeholder engagement, and problem-framing skills, with the judgment to recommend the right solution for the business problem — whether that is traditional analytics, machine learning, generative AI, workflow automation, or no AI solution at all.