AI Engineer

Amber; United States; Remote; USD 120,000 - 180,000; Full time

Benefits offered by this job

Equity

Health coverage

Dental and vision

PTO 25 days

Learning budget

Home office stipend

Quarterly team weeks

Job summary

Amber is seeking an applied software engineer to build AI assistant capabilities for a product handling live operational information. You will work across retrieval, orchestration, evaluation, and production integration to enable users to ask questions and take action within the same experience.

The role emphasizes measurable performance in real workflows, prioritizing shipping and real-world impact over polished demos. Remote within the Americas is supported along with a São Paulo base.

Qualifications

• Three+ years of production software engineering experience, including recent work with LLM applications.
• Experience with retrieval systems, structured model outputs, and evaluation tooling.
• Strong software engineering skills in Python or TypeScript.
• Good judgment about where AI is useful and where a conventional interface is more appropriate.
• Focus on shipping and measuring improvements rather than polishing demos.

Responsibilities

• Design retrieval and orchestration pipelines for product, order, supplier, and freight information.
• Create evaluation tools to monitor answer accuracy as the product changes.
• Deliver assistant-powered workflows in collaboration with product engineers.
• Balance response speed, operating cost, and output quality across model providers.
• Run fast experiments, assess results, and discontinue approaches that do not prove useful.

Skills

LLM apps

Python

TypeScript

Retrieval systems

Evaluation tooling

Rapid experimentation

AI usage judgment

Tools

Git

Docker

APIs

Job description

Role overview

Build the AI assistant capabilities for a software product that handles live operational information. The role spans retrieval, orchestration, evaluation, and production integration, enabling users to ask questions and take action through the same experience. This is an applied engineering position focused on measurable performance in real workflows.

Responsibilities

• Design retrieval and orchestration pipelines for product, order, supplier, and freight information.
• Create evaluation tools to monitor answer accuracy as the product changes.
• Deliver assistant-powered workflows in collaboration with product engineers.
• Balance response speed, operating cost, and output quality across model providers.
• Run fast experiments, assess results, and discontinue approaches that do not prove useful.

Requirements

• At least three years of production software engineering experience, including recent work with LLM applications.
• Practical experience with retrieval systems, structured model outputs, and evaluation tooling.
• Strong software engineering skills in Python or TypeScript.
• Good judgment about where AI is useful and where a conventional interface is more appropriate.
• Focus on shipping and measuring improvements rather than polishing demos.

Nice to have

• Experience building agent workflows that interact with ERP-like systems.
• Public evaluation work, open-source tools, or writing about applied LLM engineering.
• Familiarity with supply chain or commerce data.

Benefits and work setup

• Full-time role based in São Paulo or remote within the Americas.
• Salary, equity, health, dental, and vision coverage.
• 25 days of paid time off, an annual learning and conference budget, and a home office stipend.
• Quarterly team weeks in São Paulo.