Portrait

Marcel Gregoriadis

BeFoo

Startup

BeFoo was a startup in the FoodTech industry, which went bankrupt in 2021. The business of the company was to create a subscription service for the delivery of prepared meals which were optimized to the customer's food preferences, fitness goals, and ultimately nutritional needs.

I joined the team as the CTO, responsible for the realization of the IT infrastructure behind the company, including customer-facing and internal apps.

Tech Stack: Go, Postgres, AWS, GraphQL, Vue, Stripe

BeFoo

Tales of Us

Startup

Tales of Us is a startup that creates, prints, and delivers personalized storybooks with AI-generated stories and illustrations.

Tech Stack: Firebase/Firestore, Python, LangChain, Next.js, Svelte, Stripe

Tales of Us Logo

Abracadalo

Startup

Abracadalo is a premium API service to aid the development of No-Code apps, specifically targetting the platform Adalo. I built this business with one of my cousins, who is responsible for the graphic design and marketing.

It offers over 40 APIs, including string utilities, a calendar event generator, and a sophisticated Notification API supporting push notifications, SMS, and emails (including custom templates).

The project is still maintained and has an active amount of subscribed users.

Tech Stack: Firebase/Firestore, Node.js, Vue, Nuxt.js, Stripe

Abracadalo Logo Abracadalo Banner

Personality Doppelganger

Non-Profit

Personality Doppelganger is a non-profit project that I started in 2023. As someone very interested in personal psychology, and a big fan of the Big Five model, I wanted to create a website that could help people discover and connect with people who share similar personality traits as them. This project relies on a huge number of people participating, since the likelihood for such a match is very low (hence this project).

I managed to set this project up in a strictly serverless manner (i.e., at zero cost, except for the domain).

Tech Stack: Firebase/Firestore, Node.js, Svelte, Supabase, Postgres (pgvector)

Personality Doppelganger Logo

filter.js

Open Source

filter.js is a lightweight JavaScript library that is able to parse string-based but SQL-like queries and apply the semantics on JSONs.

The challenge with this was to parse the string in a way that was flexible, error-tolerant, and yet powerful, essentially programming a very high-level programming language (💁🏼‍♂️ SQL) on top of JavaScript.

Tech Stack: TypeScript

To view even more open source projects, check out my other GitHub repositories.