Neod

Inspiration

One day i was learning about neural networks then a idea arrouse in my mind that what if we have a better way to learn these then I got an idea to build this.

What it does

Helps new learners to understand what neural networks are , and what exactly then does in a AI Model, Simplies the Complexity of building a neural network by just drag and drop nodes, and making it modular for new minds.

How we built it

we used Nextjs with heroui for frontend and for nodes we used reactflow which is a library for drag and drop node based canvas , and for server we used nextjs only and for database used postgresql ,and Oauth for authentication.

Challenges we ran into

while working on the nodes we faced problem in managing too many nodes for each concept so we grouped them in categories like activation , algorithms, loss, etc , then built a generic node then used parametric functions to built according to the type of each category

Accomplishments that we're proud of

Perfectly integrated Tensorflow JS with the Nodes to handle the hyperparameters from the nodes and use those values in the tensorflow workflow to build the neural network and run the model. we were proud at that time.

What we learned

Learned many things along the way . from frontend building with managing a codebase with 20k+ lines of code.handling project managements tools, collaborating.

What's next for Neod

we are working to get most of the functions available to users and figures out to build it for pro users also , which will help them in build a chatgpt like ai model , which will feel like they are using a n8n working but from root.

Built With

• chatgpt
• copilot
• cursor
• express.js
• heroui
• javascript
• nextjs
• node.js
• postgresql
• python
• react
• reactflow
• typescript