Senior Research Engineer - Tech Focus

- Design and develop AI systems utilizing foundation models, prompt engineering, retrieval-augmented generation, multi-agent architectures, and classical machine learning techniques - Fine-tune large language models on domain-specific datasets and assess their performance utilizing offline and online techniques like A/B testing and telemetry - Transform prototypes into production-quality services applying rigorous software engineering and MLOps practices - Lead original research initiatives and produce thought leadership outputs such as whitepapers and patents; implement findings into operational capabilities - Continuously review emerging technologies; identify and adapt promising methodologies for Microsofts applications - Manage the end-to-end pipeline from data preparation, model training, and evaluation to deployment and feedback loops - Detect and address issues related to model quality, latency, and scalability using tools like PyTorch or TensorFlow - Oversee CI/CD and MLOps processes including model version management, re-training, and evaluation - Collaborate closely with engineering and product teams to integrate AI components into Microsoft offerings - Establish and maintain robust evaluation and experimentation frameworks to assess model/system performance - Create feedback loops for continuous learning based on user input and system analytics - Conduct controlled experiments, analyze outcomes, and drive data-informed decisions for product and model enhancements - Develop proofs of concept to validate innovative ideas on a practical scale - Curate valuable datasets, including synthetic data, and implement labeling and quality assurance protocols tied to KPIs - Collaborate with engineers, data scientists, designers, and product managers to create impactful AI functionalities - Transform research advancements into scalable applications that align with product goals - Clearly communicate insights and decisions in internal meetings, demonstrations, and documentation - Identify potential risks regarding fairness, privacy, safety, and data integrity; work to mitigate these risks - Adhere to Microsofts Responsible AI principles throughout the project lifecycle - Contribute to internal policies and tools promoting responsible AI practices - Read, critique, and adapt contemporary research, identifying gaps and proposing methods with defined trade-offs - Implement ideas into maintainable and efficient code, ensuring integration with CI/CD and optimizing for performance metrics - Specialize in large-scale training and optimization of various model types, including dialogue agents and multimodal systems - Address real-world data challenges such as data incompleteness, label noise, and class imbalances.