AWS Bedrock, Lambda, ECS, S3, OpenSearch, and IAM

Overview

• Bedrock provides managed access to supported foundation models and related capabilities.
• Lambda runs event-driven functions; ECS orchestrates containers; S3 stores objects such as original PDFs.
• OpenSearch provides search infrastructure, while IAM controls access to AWS resources.

Why this matters

AWS organizations can integrate inference into existing compute, storage, search, and identity operations.

An everyday example

Use the company's warehouse, processing desks, access system, and approved external specialists.

A technical example

S3 upload → event/queue → ECS parsing worker → search index → Node API → Bedrock generation.

How it works

1. Apply IAM/service identities
2. Validate upload
3. Run bounded compute
4. Index evidence
5. Retrieve with tenant filters
6. Generate
7. Validate/audit.

Architecture and flow

Web API → Lambda/ECS → S3 + OpenSearch + Bedrock, with IAM controlling service access.

The insurance assistant example

The insurer stores source documents in S3 and uses an approved model through Bedrock for grounded explanations.

When to use it

AWS is already the organization's operational home and the required model/features are available in approved regions.

When to avoid it

Do not assume all Bedrock models share tool/schema behavior or use Lambda for jobs that exceed its applicable constraints.

Interview answer

“AWS supplies model access, compute, objects, search, and service identity. I still implement tenant-level permissions and workload-specific reliability.”

Benefits, trade-offs, and production considerations

• IAM resource access and application customer authorization are different.
• Evaluate service quotas, data paths, cold starts, and container costs.
• Bedrock, Lambda, ECS, S3, IAM.