As Microsoft continues to push the boundaries of AI, we are on the lookout for individuals to work with us on the most interesting and challenging AI questions of our time. Our vision is to build systems that have true artificial intelligence across agents, applications, services, and infrastructure. We aim to make AI accessible to all , consumers, businesses, developers , so that everyone can realize its benefits.
In this role, you will design and build models and ML pipelines for user and conversation understanding , helping Copilot better interpret user intent, extract meaning from conversations, and deliver more relevant, personalized experiences. You will work across the full model lifecycle, from data preparation and training to evaluation and production deployment.
Responsibilities:
- Build and Deploy Models: Design, train, evaluate, and deploy machine learning models for natural language understanding tasks including intent detection, topic classification, conversation summarization, and user personas.
- Design ML Pipelines: Architect scalable, production-grade training and inference pipelines using Spark, Databricks, Azure ML and modern ML frameworks.
- NLP and Representation Learning: Develop and fine-tune transformer-based models and text encoders; build and maintain embedding pipelines and vector databases for semantic search and retrieval.
- Experimentation and Evaluation: Drive rigorous offline and online experimentation to measure model quality, iterate on architectures, and improve key product metrics.
- Collaborate Across Teams: Partner with data engineers, data scientists, and product teams to translate research insights into shipped features and align model outputs with product goals.
- Show Ownership Mindset: Proactively monitor model performance in production, diagnose regressions, and address scalability and reliability challenges before they become bottlenecks.
- Contribute Strategically: Identify opportunities to improve model architectures, training methodologies, and evaluation frameworks; mentor others on ML best practices.
Qualifications:
- Bachelor’s Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 4+ 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+ years related experience (e.g., statistics, predictive analytics, research) OR Doctorate in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 1+ year(s) related experience (e.g., statistics, predictive analytics, research) OR equivalent experience.
- Proven experience in NLP, including experience with modern transformer architectures for tasks such as classification, encoding, summarization, and semantic search.
- Experience with text embedding models, vector databases, and retrieval-augmented generation (RAG) patterns.
- Familiarity with distributed training, model optimization, and serving ML models at scale.
- Experience with search ranking, relevance modeling, or information retrieval systems.
- Proficiency in Python and ML frameworks such as PyTorch, Hugging Face Transformers, or similar.
- Experience working with data platforms (e.g., Spark, Databricks, Azure ML) and building end-to-end ML pipelines from data ingestion through model deployment.
Experience Level: senior Employment Type: full-time Workplace Type: onsite Category: Engineering Industry: Technology Salary Range: $119,800 – $234,700 per year Salary Min: 119800 Salary Max: 234700 Salary Currency: USD Salary Period: year Required Skills: Machine Learning, Natural Language Processing, Python, PyTorch, Hugging Face Transformers, Spark, Databricks, Azure ML Preferred Skills: Text Embedding Models, Vector Databases, Retrieval-Augmented Generation (RAG) Patterns, Distributed Training, Model Optimization, Serving ML Models at Scale, Search Ranking, Relevance Modeling, Information Retrieval Systems
XML job scraping automation by YubHub
