Staff Applied AI Engineer at Ivanti in South Jordan, Utah

Job description

• We're looking for a rare, full-stack data and AI practitioner who can operate fluidly from raw data all the way to production AI systems and the enterprise architecture that supports them. You will ingest and reason over everything from highly structured warehouse data in Snowflake to messy unstructured sources, turn it into defensible analytics, and ship agentic AI solutions that are both intelligent and ruthlessly cost-efficient.
• This is a builder-first role with real architectural ownership. You'll write the models and the agents yourself, and you'll define the reference architectures, patterns, and standards that let the rest of the organization build on top of your work. If you're equally comfortable defending a confidence interval and a system-design decision, this role is for you.
• A unified data foundation that combines Snowflake and unstructured sources for downstream modeling.
• A churn analytics product trusted by the business, with quantified uncertainty and clear recommended actions.
• At least one production agentic solution that demonstrably reduces token spend/latency versus a naive baseline while meeting quality bars.
• A documented reference architecture and set of standards adopted by other teams.

Responsibilities

• Unify structured and unstructured data. Build pipelines that pull structured data from Snowflake (and adjacent warehouses/lakes) alongside unstructured sources — text, documents, logs, transcripts — into clean, modeling-ready datasets.
• Deliver decision-grade analytics. Produce customer churn analytics with properly quantified uncertainty (confidence/credible intervals), not just point estimates, and communicate what the numbers can and can't support.
• Build predictive and prescriptive models. Move beyond "what will happen" to "what should we do about it" — forecasting, propensity, and optimization/recommendation systems that drive concrete business actions.
• Engineer agentic AI systems. Design and ship LLM-powered agents and workflows that are token-efficient by design — tight context management, retrieval and caching strategies, model routing, and evaluation harnesses that keep cost and latency low without sacrificing quality.
• Architect for the enterprise. Define reference architectures, integration patterns, and governance standards spanning data ingestion, model development, MLOps/LLMOps, security, and observability — and bring stakeholders along with clear diagrams and documentation.
• Own quality and reliability. Establish evaluation, monitoring, and guardrails for drift, accuracy, bias, safety, and cost across both classical ML and GenAI systems.
• Partner across the business. Translate ambiguous business problems into technical solutions and explain technical tradeoffs to non-technical stakeholders.

Requirements

• Strong hands-on data engineering with Snowflake (modeling, performance, cost management) and SQL, plus experience wrangling unstructured data.
• Solid applied statistics: you can build churn/retention models and correctly express uncertainty with confidence or credible intervals, and you understand the assumptions behind them.
• Demonstrated experience building predictive and prescriptive analytics that shipped and influenced decisions.
• Production experience with LLM/agentic systems — frameworks such as LangGraph, the Claude Agent SDK, CrewAI, or custom orchestrators — with a real track record of optimizing for token efficiency, cost, and latency.
• Production RAG experience (chunking, hybrid search, reranking, retrieval evals) is strongly expected at the senior+ level.
• Architecture chops: you can design and document end-to-end systems and patterns others build on, and defend those decisions with evidence.
• Strong Python and a software-engineering mindset (testing, version control, CI/CD).
• Excellent written and verbal communication; comfort working asynchronously in a distributed team.
• Cloud certifications (AWS Solutions Architect, Google Cloud Professional ML Engineer, Azure AI Engineer) and/or TOGAF for enterprise architecture.
• Experience with inference optimization (quantization, model routing, caching, vLLM/TensorRT-style serving).
• MLOps/LLMOps tooling and platform-building experience.
• Domain experience in [your industry], and prior work owning AI strategy or build-vs-buy decisions.

Skills

Required

• Snowflake
• SQL
• Data engineering
• Unstructured data wrangling
• Applied statistics
• Churn/retention modeling
• Confidence/credible intervals
• Predictive and prescriptive analytics
• LLM/agentic systems
• LangGraph
• Claude Agent SDK
• CrewAI

Preferred

• TOGAF
• Inference optimization
• Quantization
• Model routing
• Caching
• vLLM
• TensorRT
• MLOps/LLMOps tooling

Benefits

• Friendly flexible working model: Empower excellence whether you’re at home or in the office and support work-life balance.
• Competitive compensation & total rewards: Including health, wellness, and financial plans tailored for you and your family.
• Global, diverse teams:Collaborate with talented people from 23+ countries.
• Learning & development:Grow your skills with access to best-in-class learning tools and programs.
• Equity & belonging:We value every voice. Your story helps inform our solutions for a changing world.