Nameet Dutia

Graphs encode relationships. In a world where AI agents need to reason about complex systems, not just retrieve information, graph structure provides the scaffolding for causal inference, dependency tracking, and compositional reasoning. Unlike tabular or unstructured data, graphs preserve the topology of real-world systems: what connects to what and how influence flows. For networks, this is not a metaphor. Networks are natively graphs.

Amazon Leo is an initiative to launch a constellation of Low Earth Orbit satellites providing low-latency, high-speed broadband connectivity to unserved and underserved communities around the world. As a Communication Systems Research Scientist, this role owns the research and system design of the radio resource management (RRM) and radio access layers of Amazon Leo’s direct-to-device (D2D) system, delivering 3GPP-compliant service to unmodified commercial handsets.

AWS for Industries A self-learning agent does more than remember a factual answer. It persists reusable knowledge, extracts knowledge when needed, applies it to future work and produces reusable operating procedures for automation. Communication Service Providers’ (CSPs) network operations knowledge lives across data and people. Network Telemetry, alarms, topology, configuration, and incident records all comprise the Telco Data Lake.