SaaS & MVP Development

A SaaS product has to do more than work in a demo. It needs sign-up, tenants that can't see each other's data, roles, billing that matches what each plan allows, and hosting that doesn't cost a fortune before the first customer arrives. I build all of it, from the database schema to the production deploy.

PitchPool, a multi-tenant platform for competition prediction pools, shows what that looks like. It has automated scoring, live leaderboards, and subscription billing with plans, entitlement gating and webhooks behind a provider-agnostic gateway, so the payment provider can change without touching the product. Access runs on three-tier role-based access control with safeguards against privilege escalation.

Early products also need to stay cheap and stay up. On PitchPool, in-memory job scheduling and consolidated tasks cut idle database traffic by around 90%, which let serverless PostgreSQL scale to zero. On match days, BullMQ ingestion with rate limiting, circuit breaking and retries keeps the leaderboards accurate through the spikes.

I've shipped other products the same way. Cloud Prep, my own product, is a practice app for AWS certification exams with per-topic accuracy and weak-area tracking. StartDescribe is an AI SaaS for e-commerce product descriptions and images, with Stripe subscriptions, where optimised workflows raised onboarding conversion by 40%.

An MVP should be small, not throwaway. I scope the first version around the workflow your users will pay for, and build it on foundations the next features can sit on.