Amazon's Artificial General Intelligence (AGI) Team is responsibly advancing the company’s generative AI technologies, including Amazon’s most expansive multimodal Large Language Models. We are looking for a motivated Machine Learning Engineer that will join our mission by building highly scalable, resilient, and performant training code, systems and infrastructure. You will work in close collaboration with AGI engineers, applied scientists, and AWS to relentlessly pushing the limit to accelerate our mission in delivering state-of-the-art AGI models for Amazon businesses and customers.
Key job responsibilities
- Collaborate closely with applied scientists on machine learning tasks ranging from ML code & data management to training and deployment of ML models.
- Research and Develop stability and optimizations to continuously improve training KPIs, such as uptime, throughput, and good put.
- Collaborate with software engineering teams to operationalize and scale training improvements across experimentation and production workloads.
About the team
As part of the Amazon General Intelligence team, AGI Modeling Services provides training capabilities and services to accelerate invention of SoTA models across all modalities and their derivatives. These services include high-performance ML infrastructure, modeling toolkits, and optimized MLOps workflows for AGI scientists to build, train and release their models. We need your help to build the advancements required to make that a reality.
Key job responsibilities
- Collaborate closely with applied scientists on machine learning tasks ranging from ML code & data management to training and deployment of ML models.
- Research and Develop stability and optimizations to continuously improve training KPIs, such as uptime, throughput, and good put.
- Collaborate with software engineering teams to operationalize and scale training improvements across experimentation and production workloads.
About the team
As part of the Amazon General Intelligence team, AGI Modeling Services provides training capabilities and services to accelerate invention of SoTA models across all modalities and their derivatives. These services include high-performance ML infrastructure, modeling toolkits, and optimized MLOps workflows for AGI scientists to build, train and release their models. We need your help to build the advancements required to make that a reality.