Machine Learning Engineer, AGI Information - Knowledge Graphs

Amazon's AGI Information is seeking an exceptional Software Engineer to drive ML systems development in the Amazon Knowledge Graph (AKG) team. AKG is re-inventing knowledge graphs for the LLM-era, developing sophisticated ML models and pipelines that enable efficient LLM grounding and power LLM-based customer experiences.

We're looking for candidates who combine strong software engineering fundamentals with practical ML system development experience. You'll need to demonstrate expertise in building scalable, fault-tolerant distributed systems, with a track record of shipping production services that handle large-scale workloads. While ML engineering skills are important, we prioritize candidates who understand professional software engineering practices across the full development lifecycle - from system design and coding standards to testing, deployment, and operational excellence.


Key job responsibilities
- Architect AI/ML systems that power our billion-entity knowledge graph, transforming raw data into intelligent, interconnected information at scale
- Develop and optimize LLM-assisted tools that revolutionize knowledge creation, from automated ontology generation to real-time fact extraction and verification
- Design, develop and maintain ML model serving infrastructure to enable high-throughput, low-latency inference in production environments
- Collaborate with applied scientists to productionize ML models, including implementing model improvements and new architectures for knowledge mining and graph construction
- Develop efficient data processing pipelines to handle large-scale training and inference data
- Support experimentation and A/B testing infrastructure to evaluate model improvements
- Participate in code reviews, technical design discussions, and sprint planning to ensure high quality software delivery
- Strong understanding of ML fundamentals and common optimization techniques
- Experience with data processing and ETL pipelines at scale