Applied AI Engineer

UpperPlain; Time; Sur place; NOK 900 000 - 1 200 000; Plein temps

Résumé du poste

UpperPlain is seeking an AI-focused engineer to design production-grade AI workflows that turn repetitive tasks into reliable steps. You will build document intake, classification, summarization, and record-writing against client systems, and provide agents with the right tools for efficient work.

You will apply model-driven approaches, evaluate multiple providers, and keep evals close to live work, ensuring cost control and robust failure handling as the project scales.

Qualifications

• Shipping model-backed features in production.
• Built software around APIs, queues, databases, and edge cases.
• Use AI tools regularly; can identify model errors.
• Familiar with structured output, tool use, retrieval, and HITL.
• Able to work with an operations lead to identify first automation step.

Responsabilités

• Turn a repetitive operational task into a reliable AI step with clear inputs, tools, and a review path.
• Build document and email intake, classification, summarization, and record-writing against the client’s systems.
• Give agents access to the right tools: search, tickets, inventory, CRM fields, and nothing extra.
• Measure quality with real examples from the business, not only a happy-path prompt.
• Work with forward deployed engineers so the agent lives in the same platform the team already opens.
• Watch cost, evals, logs, and failure cases after launch. Production is the product.

Connaissances

Model deployment in production

APIs and data pipelines

AI tool usage & evaluation

Structured output design

Human-in-the-loop design

Outils

APIs

Queues

Databases

Description du poste

What you will do

• Turn a repetitive operational task into a reliable AI step with clear inputs, tools, and a review path.
• Build document and email intake, classification, summarization, and record-writing against the client’s systems.
• Give agents access to the right tools: search, tickets, inventory, CRM fields, and nothing extra.
• Measure quality with real examples from the business, not only a happy-path prompt.
• Work with forward deployed engineers so the agent lives in the same platform the team already opens.
• Watch cost, evals, logs, and failure cases after launch. Production is the product.

What you bring

• Hands-on experience shipping model-backed features in production, not only notebooks or demos.
• You have built software around existing systems: APIs, queues, databases, and the ugly edge cases.
• You use AI tools constantly for your own work, and you know when the model is wrong.
• Comfort with structured output, tool use, retrieval, and human-in-the-loop design.
• You can sit with an operations lead, watch them process a document, and name the step worth automating first.

AI-first, in practice

• This role is AI-first by design. You think in workflows, tools, and review, not in chatbot wrappers.
• You have used several model providers and you pick them for the job, not for the brand.
• You write evals and keep them close to the live distribution of work.