Senior Applied AI Engineer, Production Pipelines

Openkrill; United States; Remote; USD 150,000 - 230,000; Full time

Job summary

ZoomInfo is hiring a Senior Applied AI Engineer to build and own the machinery for AI-native data management. You will work on entity-agent pipelines, evaluation harnesses, and productionizing prototypes in a large-scale data environment.

You will collaborate with the Principal AI Engineer, product managers, and data teams to deploy on the Agentic Platform with robust tooling and observability. You should have shipped LLM-based systems to production, be comfortable with adapters and runners, and

Qualifications

• Shipped LLM-based systems to production with evaluation harnesses.
• Strong Python and TypeScript skills.
• Experience building platform components: adapters, runners, queues for replay/idempotency/observability.
• Understanding of data pipelines and entity resolution in production environments.
• Proven track record taking prototypes to production without losing proof points.

Responsibilities

• Run the benchmark and golden-gate system and manage versioned evaluations.
• Complete the entity-agent pipeline including enricher fan-out and location sets.
• Build evidence adapters and maintain vendor-oriented architecture.
• Deploy onto ZoomInfo Agentic Platform and ensure production readiness with rollback paths.
• Make the loop self-service with review queues and batch runners.

Skills

LLM production experience

Python

TypeScript

API design

Batch processing

Observability

Data pipelines

Tools

API layer

Batch runner

Web UI

Observability tools

Replayable store

Job description

ZoomInfo is hiring a Senior Applied AI Engineer to build and own the machinery for AI-native data management. You will work on entity-agent pipelines, evaluation harnesses, and productionizing prototypes in a large-scale data environment.

You will collaborate with the Principal AI Engineer, product managers, and data teams to deploy on the Agentic Platform with robust tooling and observability. You should have shipped LLM-based systems to production, be comfortable with adapters and runners, and