
Learn optimization strategies to reduce latency for DynamoDB workloads, wherever you’re running them
The modeling decisions you make early in a DynamoDB project — how you structure partition keys, design access patterns, and use indexes — can come back to bite you at scale.
Join DynamoDB consultant Uriel Bitton and ScyllaDB engineers for a hands-on look at how to optimize your data modeling. We’ll start by focusing on data modeling within DynamoDB itself, then discuss additional options that are available for those who have moved DynamoDB workloads to ScyllaDB (via its DynamoDB-compatible API, Alternator).
We'll cover:
- Access-pattern-driven design and why it changes how you structure your tables
- Partition key strategies that avoid hot spots and uneven throughput
- How to model your way around DynamoDB's per-partition throughput limits or avoid it with ScyllaDB Alternator
- Index design and where it hurts performance
- Working around DynamoDB limits with ScyllaDB Alternator
- Tips for hybrid vector search data modeling with ScyllaDB Alternator

Guilherme Nogueira, Technical Director
Guilherme Nogueira is an IT professional with more than 15 years of experience, specializing in large-scale database solutions. He enjoys tackling unique challenges that only NoSQL technologies can solve and helping users navigate complex and ever-scaling data environments. Outside of work, he is an avid Linux gamer. Guilherme is a technical director at ScyllaDB.

Uriel Bitton, DynamoDB Consultant
Uriel is a cloud engineer and AWS Community Builder with 15 years of experience building serverless cloud systems. He has spent much of his career working with AWS and serverless architecture, with a particular focus on DynamoDB. His work has covered everything from data modeling and access pattern design to cost optimization, performance tuning, and building databases that can scale reliably and cost efficiently.
ScyllaDB is the Monstrously
Fast + Scalable NoSQL Database
Power your applications with ultra-low database latency & extremely high throughput