Building and operating seven personal AI SaaS products on OCI
A hands-on platform case covering planning, architecture, implementation, deployment, and observability across multi-LLM, RAG, and documentation-driven AI services.
Next.jsFastAPISpring BootpgvectorQdrantTraefik
Impact
- Planned and operated seven AI SaaS and platform products end-to-end
- Runs 37+ containers across OCI ARM64 and on-prem infrastructure
- Composed a six-layer security model using Traefik, Authentik, and Cloudflare
Services
- ScrapFlow: an integrated capture platform with multi-LLM routing and semantic retrieval
- SpecOps: a team workspace for API design, review, testing, and documentation publishing
- DevFlowAI: an AI agent-driven development hub spanning requirements through deployment
- PRISM, QueryFlowAI, and Nuri SM Portal: additional products covering company generation, AI SQL workflows, and internal operations
Why It Matters
This group of services is not positioned as a set of demos. The key point is that the platform was designed as an operable system. Different technical choices, data flows, deployment concerns, auth, proxying, and security policy were unified into one operating model.
Outcome
- Validated a modern AI stack at the level of platform operations rather than isolated prototypes.
- Demonstrated both PM-grade problem framing and developer-grade implementation speed in the same body of work.
- Established a personal platform foundation that can turn new service ideas into products quickly and repeatedly.