How to apply for Quality Assurance Engineer- AI Products

About Xpansiv

Xpansiv runs the infrastructure for energy transition markets: registries, marketplaces, market execution, wholesale power, and market data for energy and environmental commodities. It has made more than 10 acquisitions since 2009 and is backed by Blackstone. The work supports auditable environmental claims, which is a concrete regulatory and market need.

About the role

You will test Xpansiv's AI products and its proprietary LLM infrastructure, covering both deterministic software behavior and non-deterministic AI outputs. The job involves writing test plans, documenting and executing test cases, and building automated and semi-automated evaluation frameworks. You act as the human-in-the-loop assurance layer for AI governance, working with AI Engineering, Product, Operations, and subject matter experts.

A typical day

You might write or update test plans and test cases for an AI-enabled product, run automated evaluation suites, and review outputs that need human judgment. You would then work with AI Engineering, Product, and subject matter experts to triage results and feed quality requirements back into development. The exact routine is not stated in the job details, so ask about team structure and release cadence in the interview.

Who Xpansiv is looking for

• Has hands-on QA experience with AI or LLM-based products, including testing non-deterministic outputs and building evaluation frameworks.
• Can design both automated and semi-automated test approaches, and knows when a human review step is required.
• Understands AI governance and safety concepts well enough to validate accuracy, reliability, and safety claims.
• Can work across AI Engineering, Product, Operations, business analysts, and subject matter experts to build quality in from the start.

Tips for this application

• Name specific evaluation methods you have used for LLM outputs, such as rubric-based scoring, golden datasets, regression suites, or human-in-the-loop review.
• Describe a time you tested a non-deterministic system and explain how you separated real defects from expected output variation.
• Connect your QA work to auditable or regulated environments, since Xpansiv's products support credible environmental claims.
• Show familiarity with energy or environmental commodity markets, or state clearly that you are learning the domain.
• Keep the application focused on testing and evaluation work, not general software engineering, because the role is QA-specific.

What to cover in your cover letter

Cover the following: your experience testing AI or LLM products; how you build automated and semi-automated evaluation frameworks; your approach to validating non-deterministic outputs; and your ability to work with AI Engineering, Product, and subject matter experts on AI governance and human-in-the-loop assurance.

Research before applying

• Read how Xpansiv describes its registries, marketplaces, market execution services, wholesale power solutions, and market data products.
• Look into what environmental commodities are and how environmental claims are audited, since the platform supports auditable claims.
• Check Xpansiv's acquisitions since 2009 to understand how the product set was assembled and where QA may be needed.
• Find out what Xpansiv has published about its AI products, LLM infrastructure, and AI governance, or prepare to ask about it.

Likely interview topics

Based on the job description, expect questions about:

• How you would write a test plan for an LLM-based product where outputs are non-deterministic.
• How to build an evaluation framework that combines automation with human-in-the-loop review.
• How to decide when an AI output is a defect versus acceptable variation.
• How you would validate accuracy, reliability, and safety requirements for an AI product under governance review.
• How you would work with AI Engineering, Product, and subject matter experts to build quality in from the start.

Common mistakes to avoid

• Treating AI QA as ordinary software QA and ignoring non-deterministic outputs, evaluation frameworks, and human-in-the-loop review.
• Claiming AI testing experience without describing concrete methods, datasets, or metrics used.
• Ignoring the domain: applying without any awareness of energy transition markets, environmental commodities, or auditable claims.