Principal Applied AI Scientist

Microsoft•Redmond, WA

About The Position

As a Principal AI Applied Scientist for Dynamics 365 Contact Center, you will play a pivotal role in advancing Microsoft's mission to empower every individual and organization on the planet to achieve more through Multimodal Agents powered by Agentic AI. You will contribute to the development and integration of cutting-edge AI technologies (Multimodal LLMs e.g. Speech, Visual, Chat – S2T, TTS etc.) into Microsoft products and services, ensuring they are inclusive, ethical, and impactful. You will collaborate across product, research and engineering teams to bring innovative solutions to life, applying your expertise in machine learning, AI applied science, & research to solve complex problems. Your work will directly influence product direction and customer experience. You will be expected to bring technical thought leadership to solve complex problems at scale. AI Mission and Impact We are in an era of unprecedented innovation and openness. As Microsoft continues to lead in AI, we are seeking individuals to help tackle some of the most exciting and meaningful challenges in the field. Our vision is to build a truly open architecture platform that enables users to summon tailored AI agents to drive real-world outcomes. We are looking for a Principal AI Applied Scientist to join our team. This role will combine AI knowledge with applied science expertise and demonstrate a growth mindset and customer empathy. Join us in shaping the future of AI agents. 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.

Requirements

Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 3+ years related experience (e.g., statistics, predictive analytics, research)
OR Master's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 3+ year(s) related experience (e.g., statistics, predictive analytics, research)
OR Doctorate in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field
OR equivalent experience.
3+ year(s) Experience in developing and deploying live production systems.
Ability to meet Microsoft, customer and/or government security screening requirements are required for this role.
Microsoft Cloud Background Check: This position will be required to pass the Microsoft Cloud background check upon hire/transfer and every two years thereafter.

Nice To Haves

Building RAGs and fine-tuning speech language models Whisper, wav2vec2 etc.
Understanding of SLU, ASR, TTS, NLP, NLU domain, Fine-tuning techniques.
Familiarity with modern LLMOps frameworks (e.g., LangChain, PromptFlow).
Experience with MLOps Workflows, including CI/CD, monitoring, and retraining pipelines.
2+ years of experience publishing in peer-reviewed venues or filing patents.
Experience presenting at conferences or industry events.
2+ years of experience conducting research in academic or industry settings.

Responsibilities

Bringing the State of the Art to Products
Build collaborative relationships with product and business groups to deliver AI-driven impact
Light weight Fine-tuning (LoRA etc.) of multimodal/MLLM models (Voice assistants etc.) to improve Agent name-entity recognition, Agent Reasoning and Agent Response quality.
Multimodal data generation, Voice assistant model evaluations - Evaluate model behavior on relevance, bias, hallucination, and response quality via offline evaluations, shadow experiments, online experiments, and ROI analysis.
Implement state-of-the-art using foundation models, prompt engineering, RAG, graphs, multi-agent architectures, as well as classical machine learning techniques.
Build rapid AI solution prototypes, contribute to production deployment of these solutions, debug production code, support MLOps/AIOps.
Ability to use data to identify gaps in AI quality, uncover insights and implement PoCs to show proof of concepts.
Leveraging Research in real-world problems
Demonstrate deep expertise in AI subfields (e.g., deep learning, Generative AI, NLP, muti-modal models) to translate cutting-edge research into practical, real-world solutions that drive product innovation and business impact.
Share insights on industry trends and applied technologies with engineering and product teams.
Formulate strategic plans that integrate state-of-the-art research to meet business goals.
Documentation
Maintain clear documentation of experiments, results, and methodologies.
Share findings through internal forums, newsletters, and demos to promote innovation and knowledge sharing
Ethics, Privacy and Security
Ensure responsible AI practices throughout the development lifecycle, from data collection to deployment and monitoring.
Contribute to internal ethics and privacy policies and ensure responsible AI practice throughout AI development cycle from data collection to model development, deployment, and monitoring.
Specialty Responsibilities
Design, develop, and integrate generative AI solutions using foundation models and more.
Deep understanding of small and large language models architecture, Deep learning, fine tuning techniques, multi-agent architectures, classical ML, and optimization techniques to adapt out-of-the-box solutions to particular business problems
Prepare and analyze data for machine learning, identifying optimal features and addressing data gaps.
Develop, train, and evaluate machine learning models and algorithms to solve complex business problems, using modern frameworks and state-of-the-art models, open-source libraries, statistical tools, and rigorous metrics
Address scalability and performance issues using large-scale computing frameworks.
Monitor model behavior, guide product monitoring and alerting, and adapt to changes in data streams.