Enterprise Healthcare AI
Private / NDAGutBot
A RAG-based conversational assistant backed by AWS Bedrock LLMs, deployed on EC2 for enterprise healthcare use.
Contact for a walkthrough ↗Developer
Role
Private / NDA
Access
01 — Problem
Enterprise healthcare teams needed a conversational assistant that answers strictly from approved source material rather than an LLM's general knowledge.
02 — Context
Built for enterprise healthcare use, where an assistant asserting something ungrounded is a liability, not just an inconvenience.
03 — System Built
A retrieval-augmented generation assistant backed by AWS Bedrock LLMs, deployed on EC2 for production use.
04 — Engineering
Architecture and deployment details are covered by an NDA with the client — get in touch if you'd like a walkthrough under confidence.
05 — Decisions
Chose AWS Bedrock to keep the LLM layer inside the client's existing AWS security boundary rather than calling an external API.
06 — Result
Deployed to EC2 for enterprise healthcare use.
07 — What I Learned
Working inside an existing AWS Bedrock deployment meant the hard problems were mostly about integration and retrieval quality, not model choice.
08 — Next Iteration
Expanding retrieval coverage as the client's approved document set grows.