AI Solutions Architect
Camille Aubert
I turn complex enterprise ecosystems into intelligent, production-grade systems — RAG, agents, and AI-assisted delivery.
/ About
Architecture is the new productivity.
For nine years I've built business software — first as a project manager who shipped code, now as the person who designs the systems and the workflows around them.
At Cegos, I lead the technical evolution of a multi-country platform ecosystem, and I've made AI a core part of how the team ships: retrieval systems that expose enterprise knowledge safely to LLMs, and automation that turns a Jira ticket into a deployed feature across repositories.
I care about the 1% — the last percent of polish, reliability, and craft that separates tools people use from tools people trust.
/ Career
A brief overview of my career
AI Solutions Architect
Cegos Group
Leading the technical evolution of the corporate multi-tenant platform ecosystem, and making AI a core part of how the team ships software.
- → Designed and shipped an enterprise RAG platform that safely exposes internal knowledge to LLM layers — built to replace the legacy enterprise search engine
- → Leading the modernization of the multi-tenant ecosystem (13 platforms across 7 countries) on AWS and Docker
- → Built the team's AI-assisted engineering workflow: ticket-driven multi-repo orchestration, AI agent procedures, structured code review
- → Driving tech-stack evolution across Laravel, React, WordPress
Project Manager & Web Developer
ASSERCAR
Full-stack delivery of business applications with strong SQL depth, API integrations, and project coordination responsibilities, including training and knowledge sharing.
AI Solutions Architect
Cegos Group
Leading the technical evolution of the corporate multi-tenant platform ecosystem, and making AI a core part of how the team ships software.
- → Designed and shipped an enterprise RAG platform that safely exposes internal knowledge to LLM layers — built to replace the legacy enterprise search engine
- → Leading the modernization of the multi-tenant ecosystem (13 platforms across 7 countries) on AWS and Docker
- → Built the team's AI-assisted engineering workflow: ticket-driven multi-repo orchestration, AI agent procedures, structured code review
- → Driving tech-stack evolution across Laravel, React, WordPress
Project Manager & Web Developer
ASSERCAR
Full-stack delivery of business applications with strong SQL depth, API integrations, and project coordination responsibilities, including training and knowledge sharing.
/ Featured Projects
A selection of professional work, published apps, and experimental prototypes.
Enterprise RAG Platform
Replaced the enterprise search engine with a retrieval-augmented platform that safely exposes internal knowledge to LLM layers — vector search pipelines, context engineering, and cost-aware inference.
Autonomous Agent Workflow & Custom RAG
Conceived and developed a production-ready autonomous agent system for multi-source indexing. Built on MCP principles to connect isolated databases with generative models, enabling reliable automated task execution.
/ Tech Stack
The tools and practices I work with, by domain
AI & Data
Backend & Architecture
Platform & Cloud
Delivery & Growth
/ Now
What I'm up to
-
Shipping an enterprise RAG platform to production — the full write-up is in Writing.
-
Building AI automation that runs delivery workflows end-to-end: ticket in, deployed feature out, across repositories.
-
Exploring multi-agent orchestration and MCP.
GitHub activity
Recently pushed · live repos
Latest writing
All posts Camille Aubert
Sep 17, 2026 · Writing
From single-task to hierarchical multithreading with AI
When an AI agent works 30 minutes on its own, waiting achieves nothing. I organize my attention into three ranked threads: prod logic, a light POC, and writing. A simple method for everyday work.
Read the post
Camille Aubert
Sep 14, 2026 · Writing
We replaced our SaaS search engine with a homegrown RAG (and AI multiplied us)
For years, search on our e-commerce platform ran on an external SaaS engine — opaque, rigid, impossible to evolve. In late 2025, I proposed a homegrown RAG engine on our existing stack. Live since late summer: roughly 47 tickets, a team of four, and a way of working that AI has profoundly changed. Here is the debrief, unfiltered.
Read the post
/ Contact Me
Feel free to reach out. I'm always open to discussing new projects and opportunities.