DESCRIPTION
About the Organization
AWS is on a mission to transform how businesses operate by delivering intelligent, cloud-powered applications. Our Applied AI Solutions organization accelerates customer success through intuitive, differentiated technology that solves enduring business challenges — blending vision with real-world expertise to build turnkey solutions that are easy to adopt and built to scale.
Within this organization, we are building the next generation of secure, intelligent workspaces — environments purpose-built for human-AI collaboration at enterprise scale.
The Role
We are looking for a Senior Applied Scientist to build the predictive intelligence powering capacity management for our workspace platform — developing machine learning systems that forecast demand, optimize resource allocation, and enable cost-efficient scaling at massive scale.
This role requires someone who can translate complex business requirements into production ML systems, designing algorithms that balance customer experience with operational efficiency across a large and diverse fleet of capacity pools.
What You'll Do
- Architect and implement ML foundations for capacity management, building models that continuously learn and optimize across multiple dimensions including geography, platform, and instance type.
- Develop demand forecasting systems that anticipate usage patterns hours to weeks in advance, enabling proactive capacity decisions at scale.
- Build anomaly detection systems that identify capacity risks before they impact customers, improving service reliability and resilience.
- Design optimization algorithms that make high-frequency, automated decisions balancing two critical forces: ensuring a flawless customer experience where every operation succeeds, while maximizing cost efficiency through intelligent resource utilization and placement strategies.
- Apply advanced ML techniques including time-series forecasting, reinforcement learning, and causal inference to measure the true impact of capacity decisions on customer experience and cost.
- Engineer features from large-scale datasets spanning usage signals, session patterns, and infrastructure telemetry — capturing complex interactions across diverse workload types.
- Partner closely with product and engineering teams to translate product vision into scientific solutions, deploying models that process millions of predictions daily with sub-second latency requirements.
What Success Looks Like
- ML systems that enable the service to remain profitable while capacity-related customer impacts become increasingly rare.
- Measurable business impact through reduced capacity waste, improved cost efficiency, and elimination of customer-impacting capacity events.
- Scientific innovation that unlocks significant cost savings through predictive resource commitment strategies and intelligent automated decision-making.
- Models that maintain the safety margins needed to absorb demand volatility without customer impact.
- An ML foundation that enables distributed, autonomous decision-making while maintaining consistent quality at scale.
What We're Looking For
- Deep expertise in machine learning, with hands-on experience building and deploying production ML systems.
- Strong background in time-series forecasting and handling demand volatility across diverse workload patterns.
- Experience with reinforcement learning for dynamic resource allocation and causal inference for impact measurement.
- Ability to work with large-scale datasets and engineer features that capture complex, multi-dimensional interactions.
- Strong systems thinking — able to design end-to-end ML pipelines that operate reliably at scale with low-latency requirements.
- Excellent collaboration skills — comfortable partnering with product managers, engineers, and business stakeholders to drive scientific solutions from concept to production.
- A track record of measurable business impact through applied ML research and deployment.
Key job responsibilities
1/ Work independently on ambiguous problems: Independently work on capacity forecasting problems that are not well defined or structured, identifying and framing new research challenges associated with broad problem areas, delivering with limited guidance.
2/ Influence across multiple teams: Drive alignment on ML approaches and capacity strategies across product, engineering, and operations teams. Actively mentor and develop others on the team.
3/ Deliver end-to-end production solutions: Develop and deliver complete solutions including scientific contributions that are deployed in production. Make technical trade-offs balancing long-term invention with short-term delivery
Lead on medium-to-large business problems: Take the lead on capacity management challenges that deliver significant benefits to customers and the business through improved forecasting accuracy and cost optimization.
4/ Drive team scientific agenda: