Available - project-based work / ~20 yrs shipping software
← Back to work
Laravel React AWS Bedrock pgvector

ASICS B2B Chat

Enterprise RAG assistant for developer support - secure, source-backed answers.

Role

Senior Full-Stack Engineer

Sector

Global enterprise

Challenge

A large enterprise needed to give developer partners fast, reliable technical answers without routing every query to a human. Generic chatbots were not an option: answers had to be accurate, grounded in the company's own documentation, traceable to a source, and secured to enterprise standards.

What I did

Designed and built a production RAG assistant on AWS, owned end-to-end: a secure ingestion and retrieval pipeline over the company's own documentation, vector search with pgvector, grounding and an evaluation step to keep answers trustworthy, a multi-turn chat experience in Laravel and React, and the AWS infrastructure (Bedrock, Postgres, Terraform) and guardrails that keep an LLM reliable in production.

The solution

A secure, multi-turn assistant returning source-backed answers from the company's own knowledge base - the hard part of RAG done properly, not a demo.

Outcome

A production AI system trusted for real developer support inside a global enterprise.

Stack
Laravel React AWS Bedrock pgvector Postgres Terraform
"Accurate, grounded in our own docs, traceable to a source - and secure enough for the enterprise."
- Let's see if we're a fit

A free, no-pressure 20-minute call.

Tell me what you're building or what's stuck, and I'll be straight about whether I'm the right person.

Book your fit call
Hiring for a role? Freelance Laravel developer Freelance Python developer