Principal Research Scientist, AI-driven Optimization

Description

As a Principal AI Scientist, Optimization, you will drive the core innovation and architectural blueprint for our intelligent services, leading a pivotal strategic initiative to transform our static optimization pipelines into a dynamic, self-learning platform. This transformation is critical to dramatically improving our operational efficiency and enabling our systems to intelligently adapt to real-time market trends. The systems you architect will manage millions of service appointments daily across thousands of our most critical global customers, powering a platform with over $1 billion in revenue and driving measurable efficiency gains that translate directly into substantial customer value.

Your primary challenge will be to integrate state of the art ML and other latest AI technologies with classical optimization, creating reliable, intelligent systems for dynamic resource allocation and scheduling. You will set the technical direction for how we research, design, and build models that must balance the competing demands of predictive accuracy, computational scalability, and real time decision making.

This is a senior leadership role focused on technical excellence and innovation. You will mentor a talented team of engineers and scientists, guide the end to end technical strategy from innovative research to production deployment and ultimately define the art of the possible for our most critical operational systems.

Responsibilities

• Lead the design and architecture of complex, end to end AI systems, guiding projects from initial conception and research through to production deployment and impact analysis.
• Own the technical vision for the next generation of intelligent automation, creating patented technologies that directly impact Salesforce Field Service's core competitive advantage and market differentiation.
• Provide technical leadership and set the strategic direction for the research and development of novel optimization and machine learning solutions.
• Drive fast paced experimentation by independently or collaboratively prototyping solutions that showcase how new AI technologies can address real world business challenges. Collaborate with cross functional teams to translate prototypes into actionable solutions and define incremental delivery plans for implementation.
• Mentor and guide a team of AI developers and scientists, fostering a culture of innovation, technical excellence, and continuous learning.
• Stay at the forefront of academic and industry advancements by attending top tier conferences, networking with experts, and continuously learning about the latest in optimization, AI, and large language models (LLMs) to identify, evaluate, and apply emerging techniques that drive a significant competitive advantage.
  
  

• Proven expertise in both classical optimization and machine learning, with a significant track record of blending solvers and models to solve complex industrial problems.
• Deep experience and theoretical knowledge in one or more of the following areas: large-scale forecasting, reinforcement learning, or data-driven decision systems.
• Hands-on proficiency in modern ML frameworks like PyTorch or TensorFlow and a strong proficiency of the data science ecosystem (e.g., Scikit-learn[sklearn]). Deep familiarity with modeling and forecasting time-series data is essential (e.g., ARIMA, Prophet, LSTMs).
• Strong understanding of probabilistic modeling, statistical evaluation, and the use of simulation-based testing to rigorously validate complex models and decision systems in an offline environment.
• Demonstrated experience architecting and deploying production ML systems. You can design robust, end-to-end pipelines—from data ingestion to model serving—and have experience with core MLOps tooling (e.g., MLflow, Airflow, Docker/Kubernetes).
• Ability to provide strong technical leadership and mentorship to a team of engineers, setting a high standard for technical excellence, innovation, and sound engineering practices.
  
  

• A Ph.D in a quantitative field such as Computer Science, Mathematics, Physics, or a related discipline.
• Advanced experience with online learning, contextual bandits, or applying Reinforcement Learning (RL) specifically to optimization problems.
• Familiarity with emerging research areas like differentiable optimization or learning-to-optimize (L2O) frameworks.
• Prior experience leading or working within cross-functional teams that include data scientists, operations research specialists, and software engineers.
• Knowledge of scalable data processing and distributed computing frameworks (Ray, Spark, Dask) and modern MLOps platforms (Weights & Biases, DVC).