Ai Research Engineer • Senior

Who you are

You have a strong background in AI and data engineering, with a focus on building scalable solutions for data quality evaluation. Your experience includes designing frameworks that assess the quality of AI training datasets, ensuring they meet high standards for compliance and performance. You are proficient in Python and have hands-on experience with Azure, which you leverage to create automated processes that enhance dataset reliability and safety.

You understand the importance of data diversity and coverage in AI model training, and you have successfully implemented methodologies that connect dataset properties to model outcomes. Your analytical skills allow you to define quality signals that align with model performance, and you are adept at developing metrics that facilitate continuous improvement in dataset quality.

You thrive in collaborative environments, working closely with cross-functional teams to establish best practices and deliver reusable tools. Your ability to communicate complex ideas clearly makes you an effective partner in running experiments and integrating evaluation processes into the data lifecycle.

Desirable

Experience with synthetic data generation and bias detection is a plus, as is familiarity with LLM-driven approaches to dataset enhancement. You are passionate about ensuring data quality and compliance, and you are always looking for innovative ways to improve existing processes.

What you'll do

In this role, you will design and implement a comprehensive data quality evaluation framework that supports the AI Data Platform team at Microsoft. Your responsibilities will include developing scalable metrics and testing methodologies that ensure the integrity of AI training datasets. You will create automated reporting systems that provide insights into dataset quality, enabling teams to make informed decisions about data usage.

You will define and operationalize quality signals that are aligned with model outcomes, focusing on aspects such as coverage, diversity, and labeling consistency. Your work will involve developing task- and model-aware evaluation approaches that connect dataset properties to training performance, reliability, and safety.

Additionally, you will build guardrails for synthetic data, implementing filtering, scoring, and provenance tracking mechanisms to ensure compliance and quality. You will collaborate with engineering teams to integrate your evaluation and generation processes into the platform's end-to-end data lifecycle, ensuring that datasets are continuously iterated upon and improved.

What we offer

Microsoft offers a dynamic work environment where innovation is encouraged. You will have the opportunity to work on cutting-edge AI technologies and contribute to projects that have a significant impact on the company's AI capabilities. We provide competitive compensation and benefits, along with opportunities for professional growth and development within the organization. Join us to be part of a team that is shaping the future of AI at Microsoft.