Overview

You’re joining Core AI, the team at the forefront of redefining how software is built and experienced. We create the foundational platforms, services, and developer experiences that power next-generation applications using Generative AI, enabling developers and enterprises to unlock the full potential of AI to build intelligent, adaptive, and transformative software.

You will be a technical contributor driving the applied science foundation for observability in AI agents and multi-agent systems running at scale. This role focuses on understanding how intelligent agents behave in production—their quality, safety, reliability, cost, and evolution over time. You will develop and apply scientific methods, evaluation frameworks, and measurement systems that help teams understand, benchmark, diagnose, and safely improve agent-based systems with confidence.

AI agents introduce fundamentally new observability challenges: non-deterministic execution, tool- and model-driven decision paths, emergent multi-agent behaviors, and quality signals that go far beyond traditional uptime metrics. In this role, you will operate at the intersection of agent architecture, telemetry, evaluation science, and responsible AI, shaping how Microsoft measures and improves observable AI systems.

Microsoft’s mission is to empower every person and every organization on the planet to achieve more. As employees we come together with a growth mindset, innovate to empower others, and collaborate to realize our shared goals. Each day we build on our values of respect, integrity, and accountability to create a culture of inclusion where everyone can thrive at work and beyond. In alignment with our Microsoft values, we are committed to cultivating an inclusive work environment for all employees to positively impact our culture every day.

Responsibilities:

What You’ll Do

• Develop evaluation and measurement frameworks for single-agent and multi-agent systems, spanning quality, safety, reliability, cost, and behavioral consistency.

• Design methodologies that connect offline evals, online signals, and production telemetry to explain how prompt, tool, model, or orchestration changes affect real-world agent performance.

• Define scientifically grounded quality signals and benchmarks for agent systems, including task success, tool-use effectiveness, plan quality, failure modes, coordination quality, and user-perceived outcomes.

• Build models and analysis techniques that help detect regressions, identify root causes, and characterize agent behavior across diverse workflows and environments.

• Advance observability for AI systems through new approaches to trace analysis, agent health modeling, behavioral clustering, anomaly detection, and multi-agent coordination analysis.

• Partner with engineering teams to operationalize evaluation and observability methods in production systems, enabling safe iteration through staged rollouts, experimentation, A/B testing, and automated regression detection.

• Contribute to instrumentation and semantic standards for agent observability, helping make agent execution more explainable, diagnosable, and comparable across systems.

• Collaborate deeply with product and platform teams across Foundry, Azure Monitor, and agent runtimes to shape end-to-end experiences for evaluation, benchmarking, monitoring, and investigation.

• Act as a technical leader by setting scientific direction, driving research-informed product decisions, mentoring others, and raising the technical bar across the organization.

Technical Focus Areas

• Evaluation science for agent and multi-agent systems: offline, online, and continuous evals; benchmark design; synthetic data; task success measurement

• Agent and multi-agent architectures: planners, tool use, memory, orchestration, and coordination patterns

• Applied machine learning and statistical methods for behavioral analysis, anomaly detection, experimentation, and regression detection

• Observability data for AI systems: traces, logs, metrics, evaluations, and cost/performance signals

• Safety and responsible AI signals: policy compliance, risk detection, auditability, and safe logging

• Benchmarking and experimentation for agent systems, including A/B tests, canaries, and staged rollouts

• Explainability and diagnosis for complex agent workflows and model-driven decision paths

Qualifications:

Required Qualifications:

• Bachelor's Degree in Computer Science or related technical field AND 6+ years technical engineering experience with coding in languages including, but not limited to, C, C++, C#, Java, JavaScript, or PythonOR equivalent experience.

Other Requirements:

Ability to meet Microsoft, customer and/or government security screening requirements are required for this role. These requirements include, but are not limited to the following specialized security screenings:

• Microsoft Cloud Background Check: This position will be required to pass the Microsoft Cloud background check upon hire/transfer and every two years thereafter.

Preferred Qualifications:

• 6+ years of experience in applied science, machine learning, evaluation systems, or related technical fields

• Strong experience designing evaluation methodologies, experiments, or measurement systems for complex intelligent or distributed systems

• Experience analyzing large-scale production or experimental data to derive actionable insights and drive product or system improvements

• Strong coding and prototyping skills in Python or similar languages, with the ability to work closely with engineering teams on production-facing systems

• Demonstrated ability to lead cross-team technical direction through scientific depth, influence, and strong problem framing

• Advanced degree in Computer Science, Machine Learning, Statistics, Applied Mathematics, or related field

• Experience building or evaluating LLM- or agent-based systems in production

• Familiarity with agent frameworks such as Lang Chain, Lang Graph, OpenAI SDK, or equivalent orchestration frameworks

• Experience with evaluation frameworks for AI systems, including benchmarking, regression analysis, and human-in-the-loop assessment

• Experience with observability systems, telemetry analysis, or distributed tracing data in large-scale environments

• Background in AI safety, guardrails, and responsible AI measurement

• Experience with experimentation platforms, causal inference, or statistical methods for product and model evaluation

• Experience working with cloud-scale monitoring platforms such as Azure Monitor / Application Insights or equivalent

Applied Sciences IC6 - The typical base pay range for this role across the U.S. is USD $163,000 - $296,400 per year. There is a different range applicable to specific work locations, within the San Francisco Bay area and New York City metropolitan area, and the base pay range for this role in those locations is USD $220,800 - $331,200 per year.

Certain roles may be eligible for benefits and other compensation. Find additional benefits and pay information here:
https://careers.microsoft.com/us/en/us-corporate-pay

This position will be open for a minimum of 5 days, with applications accepted on an ongoing basis until the position is filled.

Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance with religious accommodations and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.