— Stack
The stack we work with.
One page per service: how it fits in a platform, the choices that drive cost, and what we see in production.
DataAmazon S3The storage layer under a data platform. Everything writes to it and reads from it, and the bill comes less from how much you store than from how you store it and how often you touch it.Read the page →DataS3 TablesManaged Apache Iceberg tables in a new S3 bucket type. You store tabular data as Iceberg tables and AWS runs the table upkeep for you. Three choices decide how it connects and what it costs.Read the page →DataAWS GlueThree things under one name: a catalog every engine reads, crawlers that fill it, and serverless Spark jobs. Compute is billed by the second, the catalog by the object.Read the page →DataAmazon AthenaServerless SQL over files in S3. You pay for the bytes scanned, so the way your files are laid out decides most of the bill.Read the page →DataAmazon RedshiftA columnar cloud data warehouse. You load data once into managed storage and query it many times, and two ways of running it decide how you pay.Read the page →DataAmazon EMRManaged Spark, Hive, Trino, Flink and HBase on open-source versions you pick. How you run it, on clusters or serverless, decides how you pay.Read the page →DataLake FormationAccess control for the data lake, written as grants on the catalog instead of IAM policies on S3 paths. Lake Formation itself is free.Read the page →DataAmazon SageMaker Unified StudioThe data and AI workspace of the next-generation Amazon SageMaker. One catalog for data on AWS, Snowflake and Databricks; domains and projects so each team owns and shares its data; blueprints so the platform team hands out governed environments; and a Data Agent to query it all in plain language.Read the page →StreamingManaged FlinkManaged Apache Flink for stateful stream processing: windows, joins, enrichment and exactly-once delivery that a per-record function cannot hold. You write the job in Java, Python or SQL; AWS runs the Flink cluster, scales it, and keeps the state. Capacity is measured in KPUs, one vCPU and 4 GB each.Read the page →StreamingAmazon MSKApache Kafka as a managed service. You keep the Kafka API, clients and ecosystem; AWS runs the brokers, patching and storage. Producers write events, consumers read them, and MSK keeps the log. How you run it, on provisioned brokers or serverless, decides how you pay.Read the page →StreamingAmazon KinesisThe AWS-native event stream. Producers write records; the stream holds them in shards for 24 hours to 365 days; many consumers read the same records independently, each at their own position. You run it in on-demand mode, where AWS scales the shards, or provisioned mode, where you set the shard count yourself.Read the page →StreamingData FirehoseA managed pipe from a stream to a destination, with no code to run. You point producers at a Firehose stream; it buffers, can transform and reformat, and delivers to S3, a warehouse, a search cluster or an HTTP endpoint. You pay per GB ingested. It moves and lands data; it does not do stateful stream processing.Read the page →AIAmazon BedrockOne API to call foundation models from several providers, open-weight models included. You pay per token, and three choices decide most of the bill and of the compliance story.Read the page →AIAgentCoreBuilding blocks to run agents in production, from any framework and any model. Each piece is adopted on its own and billed on what it uses.Read the page →AIRAG on AWS with Bedrock Knowledge BasesThe RAG building block on AWS. A knowledge base ingests your documents, parses, chunks and embeds them, then returns the right passages to your LLM or agent. Bedrock can run the whole pipeline for you, or you run it on the vector store you choose.Read the page →IntegrationsDatabricks on AWSA lakehouse platform on AWS. Data comes in through managed connectors, files or streams, compute runs in your VPC or in Databricks' account, and agents reach governed data through MCP. Tables sit in your S3 unless you choose Databricks default storage.Read the page →IntegrationsSnowflake on AWSA managed warehouse that runs in Snowflake's own account inside an AWS region. Data reaches it from your AWS account in four ways, and agents on AWS can query it through MCP. You pay for compute in credits and for the storage Snowflake holds.Read the page →Integrationsdbt on AWSThe transformation layer of a data platform. Your SQL transformations become versioned, tested code that runs on the engine you already use. dbt Core is open source; the dbt platform adds a hosted scheduler, IDE and CI for a licence.Read the page →IntegrationsDuckDB on AWSAn open-source analytical database that runs inside your own process. On AWS it turns a Lambda function or a Fargate task into an ETL job: read raw files from S3 with SQL, transform them, and write Iceberg tables to S3 Tables. You pay for the seconds of compute and the storage, with no cluster and no engine licence.Read the page →