I'm absolutely thrilled to be part of ModusHack! I've been having a blast experimenting with Modus CLI and AssemblyScript, as I shared in my Kick Off article of this series. But here's the exciting part—I'm still on the hunt for that perfect project idea!

I understand there are a few methods for getting a project idea:

1. Identify “real-life” problems.

2. Search on the internet.

3. Read 100 research papers to apply one yourself. Easy stuff!

However, that approach hasn't quite worked for me in any hackathon. The challenge with ideas is that some are very common (like todo apps), some are repetitive, and some feel just out of my reach.

I decided to do what I do best.

Research.

All you need to do is

• Read blog posts on Hypermode’s Official Blog.

• Read recent blog posts on Hashnode’s Official Blog.

• Watch Youtube videos on Modus posted by Hypermode and Hashnode team.

• Go through the projects winners of last three hackathons hosted by Hashnode.

Right! It’s simple! :/

Hypermode Blogs

Blog posts are an excellent resource for understanding how product decisions are made. They provide insights into future plans and reveal how engineering teams design and develop the robust systems we rely on.

I could categorise the blog posts on Hypermode into three categories:

1. Product launches and guides

2. Introducing new team members

3. Case studies

I filtered the blog posts according to the categories mentioned above. I chose to skip the posts about team members and focused on reading those related to case studies and guides.

After exploring the relevant blog posts and watching the YouTube videos, I discovered that Hypermode’s team has been doing amazing work on Vector Search (Semantic Search) and AI models. The team is making customer applications faster and more efficient! I also stumbled upon the E-commerce site they've developed to showcase how to use vector search. And guess what? It's open source! This means I can dive into the repository anytime to see how Vector Search is implemented. How exciting is that?

Previous Hashnode Hackathon Winners

As Hashnode is hosting ModusHack, I thought it best to see what projects had been developed in previous hackathons. There are three major reasons:

1. Get a clear overview of recent projects and the technologies used.

2. Understand how the winners write their blog posts, including the details and explanations of their projects.

3. Identify the patterns judges look for in the projects.

I visited the Hashnode Hackathons website and selected three hackathons that are closely related to ModusHack in terms of technology, such as API and GraphQL, and focus areas like AI, vector search, and data-intensive applications.

I primarily focused on AI for Tomorrow, Hashnode APIs Hackathon and MindsDB Hackathon.

To locate the projects submitted for the respective hackathons and identify the winners, I used the specific hackathon tags (#AIForTomorrow, #MindsDB, etc.) on Hashnode. This search led me to the winners' announcement posts and other projects submitted for the hackathons. I shortlisted the projects based on tech stacks and which I personally found interesting.

For reviewing these projects and maximum learning I set a few parameters.

• Deliverables : What is the end product of the project which the users will use. Web app? API? CLI ?

• Tech Stack : Which technology stack have been used?

• Architecture and Design : How have they designed their projects

• Blog post : Is the blog post easy to understand ? Does it explains the project well?

1. AI For Tomorrow : Sighted

• Deliverables : Web site and Python SDK

• Tech Stack : Python, Nextjs, Aceternity UI, Shadcn UI and Vercel.

• Repo : https://github.com/lnxpy/sighted

• Domain : AI

Learnings

1. This project effectively addresses the challenge of reading comprehension, especially for individuals with ADHD, but it benefits everyone.

2. It integrates seamlessly into your own project. The Python SDK is customizable with templates, enabling other developers to build applications with it.

3. The blog post clearly details the project's development process and includes code snippets and figures for enhanced understanding.

In short, the project solves a real problem, can be extensively used by other developers and customised easily.

2. AI For Tomorrow : Genkin

• Deliverables : web app

• Tech Stack : Supabase, Nextjs, Gemini API

• Domain : AI / Chatbots in Finance

Learnings

1. Niched to Finance, the core idea of the project is too make handling daily finances by simply chatting with AI. No need to manually write expenditures.

2. While we are very familiar with text-based AI bots, this project is out of the box. Having an interactive chat bot which replies with tables, charts and buttons is very rare.

3. It’s not very complex.

4. Rather than having multiple features, it focuses on one which works fantastically.

3. Hashnode APIs Hackathon : Pinecone x Hashnode Semantic Search

• Deliverables : Search API

• Tech Stack : AWS Lambda, Semantic Search, Vector Database, Open AI

• Repo : https://github.com/dreamorosi/serverless-semantic-search-hashnode

Learnings

1. The project is very complex (at least to me), uses vector DB, embeddings and AWS lambda to search blog posts on Hashnode using Semantic Search (NLP)

2. The blog post is really detailed and explains how the project is build.

3. The project fits right into the Hashnode’s ecosystem.

4. Demo video gives clear explaination of the project and how it. can be used.

4. Hashnode APIs Hackathon : Hashmind

• Deliverables : web app

• Tech Stack : Nextjs, Hashnode API, Notion API, AI

• Repo : https://github.com/benrobo/hashmind

Learnings

1. The project revolves around the Hashnode’s ecosystem, particularly writing and publishing blog posts.

2. It makes writing easy.

3. Fit’s right into your life by integrating with Notion and Hashnode.

5. MindsDB Hackathon : SkillSnap

• Deliverables : Web app

• Tech Stack : MindsDB

Learnings

1. The blog post is highly detailed. It uses code snippets and demo videos extensively to clearly explain each feature of the application.

2. The idea is fairly common but they have developed every single feature and showcased used of MindsDB

3. They have tried to automate multiple parts in Hiring Process and integrated with AI models.

Hypermode Youtube Videos

Reading multiple articles and docs can be tough, but Hypermode's YouTube videos make it a breeze! You get a direct visual explanation from their team, and it totally boosts your confidence!

1. Getting Started with Modus

This video is an absolute game-changer! It explains all the amazing features of Modus in just 5 minutes! From installation to integrating AI models, the host quickly walks you through the core features of Modus, API generation, AI models, and so much more. Get ready to be blown away!

Learnings

1. The video showcases how to get going with Modus in minutes.

2. Quotes API generation and testing using Postman

3. Integrating Llama 3.1 model

2. Introducing ModusHack: Modus 101

I watched this livestream live. This live stream provides a more in-depth explaination of Modus and its features. I won’t repeat again.

The best of the video was when the host asked the Hypermode’s team member for suggesting some Hackathon Ideas:

1. CmdK : Searching globally using Cmd + K

2. Building entire application using Modus and deploying it.

3. Using vector search to replace Algolia Search

Key Takeaways

After reading and watching all the resources above, I learned so much! I discovered exciting new projects, cutting-edge technologies, and more. I also found out how to write blog posts for submitting the final project.

Have I found the Idea for my project?

Not specifically.

But the research surely made my brain think about problems, loop holes and ideas in different aspects which will eventually land me on an amazing idea. Submitting a project for a Hackathon is easy, but for winning I need to keep in mind the following:

1. The deliverables of the project are crucial. Projects must be easy to set up and understand.

2. The idea should either “solve a problem” or “give pleasure” in existing solution.

3. Demo videos are crucial to showcase the project to new pair of eyes with a visual presentation. While blog post covers in-depth explanation.

4. Specially for ModusHack, if the project includes an optimal use case of AI models and Vector Search, it would really stand out. (I lost the count the number of times I read “AI models” and “Vector Search” in those blogs, its too much).

Fin

So I read documentation, blog posts watched multiple videos but I haven’t really gotten any idea for my project? Or have I?

The only way to find out is to read the next blog post in the series.

If you really reached upto here without skipping between the lines, I appreciate it. I really do.
Thanks!

Let’s connect:

• Website : https://mahendradani.vercel.app

• Github : https://github.com/MahendraDani

• Linkedin : https://linkedin.com/in/mahendra-dani