Danil Nagy

SmartplanAI is a SaaS web app for automated office space planning. The app consists of two parts:

1. A web app built with Vue and Nuxt. The app is deployed on Netlify and uses Netlify Identity to manage users, along with a custom Auth UI. The app is mostly front end and has no dedicated server. Instead, it uses a combination of AWS Lambda, Amazon S3, and FaunaDB to manage the back-end data layer. The app also includes a custom CAD floor plan drawing and visualization tool built using Paper.js and HTML Canvas.

2. A cloud service built with Python and Flask deployed on Heroku. This service generates floor plans for any floorplan drawn in the app's UI. The plan generation functions rely on a custom polyline-based geometry kernel and graph-based spatial data structure. It also uses advanced methods of optimization and machine learning to boost the efficiency and quality of generated plans.

I developed the initial MVP of the SmartplanAI app, working as a solo developer over a six-month period. Within three months of launch, the app generated contracts over $500,000 in recurring annual revenue, leading to a successful seed funding round.

An experimental communication app built as a hobby project. Unlike most other social media and chat apps, which aim to amass contacts and push content to everyone, Tincann.ing creates a single dedicated channel for communication with one person at a time.

The app is built using a modern JAMStack low-backed approach. The front end is developed with the SvelteKit framework and deployed using Netlify. Firebase's Firestore is used to manage all the app's data, and its real-time capabilities power the app's chat functionality.

An online training, tutorial, and resource site focused on teaching computational methods and tools to architects and designers.

The site was built from scratch using React and Next.js and includes text, streamable video content, and downloadable files. To manage the content, a custom CMS was developed based on simple text markdown files.

The site includes a custom authentication system and integration with Stripe checkout to handle subscriptions. It also has a custom gating and permissions system that blocks certain content based on membership level.

A SaaS platform for analyzing, optimizing, and managing commercial office spaces. Dojo's main product is a web-based app that allows facility managers and HR specialists to manage their company's office spaces using various tools. For example, facility managers can test different variations of office layouts, while HR specialists can use the app to create employee seating arrangements. The Dojo app uses spatial analytics to analyze different layout variations to allow users to make informed decisions. It also has automation tools to help users lay out their spaces quickly and efficiently.

I was a founding member of Dojo.co when it was spun out as a technology company from parent company Silverstein Properties, one of the largest real estate development and management companies in New York City. While at Dojo, I acted as the lead developer and product manager, supervising the development of the product's prototypes and MVPs and managing a small team of freelance front-end and back-end developers. I also built all the product's spatial analytics and automation aspects using custom algorithms developed in Python.

An open-source desktop app and plugin developed using web technologies. It has a back end written in Python and Flask, which implements an optimization process based on a custom-coded genetic algorithm. It also has a front end written in Angular that runs in a browser and allows users to explore their optimizations. A custom plugin for Rhinoceros 3D/Grasshopper allows Discover to connect to the user's computational design models to run optimizations and derive the best solutions based on a set of exposed input parameters and outputs in the form of objectives and constraints.

I started developing Discover in 2018 as a personal project to assist with the teaching and consulting projects I was working on. I coded the entire Python back end myself and the genetic algorithm implementation. I also built the Grasshopper plugin using C#. I built the initial front-end prototypes using pure JavaScript but hired a freelance to rebuild the front end using Angular. I am now working on a pure desktop version of Discover using Electron and a rewritten version of the back end in Node.js.