Job Specifications
Please Note: This position is open only to candidates authorized to work in the U.S. without the need for current or future visa sponsorship. Additionally, this position is based in the Kansas City area, and we are only considering candidates who reside locally.**
At Sunlighten, we're not just about infrared saunas, we're on a mission to improve lives through innovative health and wellness solutions. As a global leader in infrared sauna therapy, we are rapidly expanding and need a talented Data Scientist, AI & BI to drive measurable impact across Sales, Marketing, CX, and Operations through applied ML/LLMs, experimentation, and analytics. This is an AI-first role: you will own evaluation, monitoring, and continuous improvement for AI agents and RAG experiences, and partner with our AI Applications Engineer to productionize safe, reliable workflows that are accurate, secure, and ROI-positive. You will also lead core BI data science work (forecasting, scoring, experimentation) and step into BI analytics/dashboarding as needed to ensure business priorities ship end to end.
Celebrating 25 years of innovation, Sunlighten has grown from its Kansas City roots to establish a global footprint, including expansion into the UK. With the global wellness market projected to reach $7 trillion in 2026, we are proud to be part of this dynamic and holistic shift. As leaders in light science and longevity, we create innovative solutions that help customers lead vibrant, active lifestyles.
Duties/Responsibilities:
LLM / Agent Quality (Applied AI)
Define evaluation strategy for LLM/RAG/agents: grounded, helpfulness, safety, regression tests, and release gates
Build and maintain "golden sets" and rubric-based scoring for agent behavior across key use cases
Establish monitoring for agent outcomes: quality, latency, cost, drift, user feedback, and business KPIs
Partner with the AI Applications Engineer on prompt strategy, retrieval patterns, tool-use behavior, and safe fallbacks
Run red team/adversarial testing and coordinate mitigations for unsafe or ungrounded behavior
Ensure privacy/security by design: PII minimization, RBAC/least privilege, secrets via Key Vault/1Password, auditable deletions (≤7 days where applicable)
Define human in the loop workflows when needed (sampling, review queues, labeling guidelines, escalation paths)
LLMOps / Governance (Production Readiness)
Own an LLM release process: prompt/model/versioning, offline evals, staging, canary, and rollback
Maintain documentation for production AI: evaluation reports, model/prompt "cards," known failure modes, and mitigation playbooks
Implement automated regression checks (pre/post deploy) to prevent quality/safety backslides
Define incident response expectations for agent issues: triage, root cause analysis, corrective actions, and follow-up measurement
BI + Applied ML (Core Data Science)
Partner with stakeholders (Sales/Marketing/CX/Ops) to convert questions into testable plans, success metrics, and decision ready recommendations
Own predictive modeling for BI priorities: lead/opportunity scoring, demand planning, end-to-end forecasting, and product/website models as needed
Design and run experiments (A/B, holdouts, quasi experimental when needed): power, guardrails, instrumentation, readouts
Define business + model metrics; build golden labels/holdouts; quantify ROI and operationalize decision thresholds
Feature engineering across Salesforce, NetSuite, Five9, Marketing Cloud, Shopify, GA4, and product telemetry; collaborate with Data Engineering to productionize in Microsoft Fabric
Translate modeling outputs into operational workflows (e.g., Salesforce scoring, routing, prioritization, dashboards, and alerts)
This is an AI-first role, but you're expected to pitch in on BI/analytics work when priorities demand it (metric definitions, semantic model alignment, dashboards, and executive readouts) to ensure outcomes land not just models
Improve data clarity: metric definitions, data quality checks, lineage notes, and stakeholder enablement
Other duties as discussed and assigned.
Requirements
2-6 years of enterprise level experience in applied data science or analytics with stakeholder-facing delivery
Bachelors or Masters degree in Data Science, Computer Science, Statistics, Operations Research (or equivalent practical experience); portfolio, GitHub or examples of shipped work preferred
Strong proficiency in Python (pandas/sklearn) and SQL, with solid statistical and experimental foundations (forecasting, power analysis, common tests)
Experience shipping models or analytics into production business workflows (e.g., CRM scoring, operational forecasting, dashboards)
Familiarity with LLM concepts (prompting, retrieval, evals) and a quality-first mindset
Working knowledge of MLOps/LLMOps and Git-based workflows, including versioning, automated eval/regression testing, monitoring/alerting, documentation, and rollback strategies.
Nice To Have (Preferred Experien