NextStep — A Student Opportunity Finder by Hari
NextStep — A Student Opportunity Finder by Hari
Finding opportunities isn't the problem. Finding the right ones is.
Have you ever searched for an internship or hackathon and ended up opening ten, twenty, or even more websites?
You find something that looks interesting.
Then you read the requirements.
“Third-year students only.”
“Must know Python.”
“Requires JavaScript.”
“Only for students from certain branches.”
And suddenly, the opportunity that looked perfect isn't actually meant for you.
Then you start searching again.
That was the problem I wanted to solve.
So I started building NextStep.
The idea behind NextStep
The idea is actually quite simple.
Instead of making a student search through hundreds of opportunities themselves, what if the student could simply tell the system:
“I'm a second-year CSE student. I know Python, React and basic machine learning. I'm looking for an internship.”
And NextStep could help find opportunities that make sense for that student.
Not just opportunities that contain the word “Python.”
Opportunities that actually relate to the student's year, branch, skills, interests and goals.
That's where the name came from.
NextStep is about helping a student figure out what their next opportunity could be.
Phase 1 — Starting with the problem
I didn't start by thinking:
“I need to build an AI application.”
I started with a much simpler question:
“Why is it so difficult for students to find opportunities that are actually relevant to them?”
There are already plenty of websites listing internships, hackathons, competitions and other opportunities.
The problem is that the student has to do all the work.
Search.
Open.
Read.
Check eligibility.
Compare.
Reject.
Search again.
I wanted to reduce that effort.
Phase 2 — Building the first version
I started with the basic student experience.
A student should be able to:
Tell NextStep about themselves.
Search for something they are interested in.
See relevant opportunities.
Understand why an opportunity matches them.
Know whether they are directly eligible or may need to build some additional skills.
I deliberately started small.
The goal wasn't to build the biggest opportunity website on the internet.
The goal was to make the basic idea actually work.
Phase 3 — Teaching the system about opportunities
There was another problem.
Human beings can read an opportunity description and understand it easily.
A computer needs information to be organized.
For example, when we see:
“Open to second and third-year computer science students with Python experience.”
we naturally understand:
Second or third year
Computer Science
Python
So I structured opportunities around information such as:
What the opportunity is
Who is organizing it
What skills are required
Which students can apply
What year they should be in
Where it is happening
When the deadline is
Where the official information can be found
This gave NextStep something it could actually work with.
Phase 4 — Making search feel more human
Students don't always type things perfectly.
Someone might type:
cpp internship
Another person might type:
C++ internship
Someone else might type:
c plus plus internship
A human understands that all three mean roughly the same thing.
So I made NextStep recognize common variations.
The same idea applies to things like:
js→ JavaScriptnodejs→ Node.jsml→ Machine Learningai→ Artificial Intelligencedsa→ Data Structures and Algorithmsweb dev→ Web Development
This means students don't have to learn a special way of talking to the application.
They can search more naturally.
Phase 5 — The question that mattered most: “Can I apply?”
Finding an opportunity is only half the problem.
The next question is:
“Am I actually eligible?”
So I built a system that checks the important requirements.
For example, if an opportunity is only for second-year students and the student is in their second year, that requirement can be matched.
If the opportunity requires a particular branch, that can also be checked.
If certain skills are required, those can be compared with the student's profile.
But I didn't want to stop there.
Because I noticed something interesting.
Phase 6 — What if a student is close, but not an exact match?
Imagine an opportunity asks for:
Python + JavaScript + Node.js
A student might know:
C++ + DSA + OOP
A normal system might simply say:
❌ Not eligible.
But is that really the whole story?
The student already understands programming.
They understand logic.
They understand data structures.
They understand object-oriented programming.
They may simply need to learn some new technologies.
This led to one of the most important ideas in NextStep:
Direct Match and Potential Match
Direct Match
The student meets the relevant stated requirements.
Potential Match
The student doesn't completely match the listed skills, but has existing knowledge that could transfer to the missing skills.
For example:
“You don't currently list JavaScript, but your C++, OOP and programming experience give you a foundation that can help you learn it.”
This doesn't mean NextStep pretends the student is officially eligible.
The official requirements still matter.
Instead, NextStep is saying:
“You may not be a direct match today, but don't automatically count yourself out.”
I really liked this idea because it changes the experience from simply rejecting students to helping them understand where they stand.
Phase 7 — Helping students see what they could become eligible for
This naturally led to another feature.
Suppose a student keeps finding opportunities that require a skill they don't have.
Instead of simply showing:
“Missing skill: Java”
NextStep can help the student understand the gap.
The student might already have related knowledge.
So the system can explain the connection between what they know and what they are missing.
This became the idea behind:
Unlock More Opportunities
The purpose isn't to tell students that they aren't good enough.
It is to show them:
“Here is what you already know. Here is what you're missing. And here is how that missing skill could open more doors.”
Phase 8 — Where AI comes in
At this point, I started thinking about where AI could genuinely help.
I didn't want to put AI everywhere just because the project was supposed to use AI.
Instead, I gave AI the jobs it is good at.
For example, a student might write:
“I'm a second-year student who likes AI and web development. Find me some hackathons.”
That's not a perfectly structured database query.
AI can help understand what the student actually means.
It can help interpret natural language and explain why certain opportunities are relevant.
But I kept an important rule:
AI helps understand and explain. The actual eligibility rules remain structured.
In simple terms:
AI can help you understand the opportunity.
The rules help determine the requirements.
Phase 9 — Making the information trustworthy
The early version of NextStep used sample opportunities while I was building and testing the application.
That was useful for development.
But a real student can't rely on made-up opportunities or broken links.
So I started moving toward a smaller collection of opportunities from reliable and official sources.
I chose to focus on quality rather than trying to collect everything on the internet.
A small number of trustworthy opportunities is more useful than a huge list filled with outdated or questionable information.
Phase 10 — Trying to take NextStep to the cloud
Once the application started coming together, I wanted to make it accessible beyond my own computer.
That brought me to AWS.
This part was definitely a different kind of challenge.
Building the application and putting the application online are two very different things.
I worked with AWS services and deployment tools to prepare the application for the cloud.
There were also plenty of practical problems along the way.
Installing tools.
Setting up accounts.
Configuring the environment.
Dealing with Docker.
Checking AWS credentials.
Fixing things after my computer setup changed.
At one point, even my development environment was affected after changing my computer's hardware.
It was a good reminder that building a project isn't always a straight line from:
Idea → Code → Finished.
Sometimes it looks more like:
Idea → Code → Error → Fix → Another Error → Search → Fix → Test → Repeat.
And that's normal.
Phase 11 — Making it feel like a real product
Once the main functionality was working, I started looking at the application differently.
Not as a collection of features.
But as something a student would actually use.
That meant thinking about things like:
Is it easy to understand?
Can someone find an opportunity quickly?
Are the important details visible?
Can the user tell why something matched?
Is it clear when someone is only a Potential Match?
What happens when there are no results?
Does it work well on a smaller screen?
Are the links actually useful?
These things may sound small.
But they make a big difference between something that technically works and something that feels like a product.
What I learned while building NextStep
The biggest thing I learned wasn't a programming language or an AWS service.
It was that building a product is mostly about solving problems one step at a time.
I started with one problem.
Then I found another.
Then another.
And each phase changed the project a little.
I also learned that AI doesn't have to replace everything.
Sometimes the best use of AI is simply helping a person communicate naturally with a computer.
A student shouldn't have to understand how a database works just to find an internship.
They should be able to say what they want.
The system should do the hard part.
What NextStep means to me
When I started this project, I was thinking about students who spend hours searching for opportunities.
But while building it, I realized there was another problem.
Sometimes students don't know what they are capable of.
They see an opportunity asking for a skill they don't have and immediately move on.
But maybe they already have five related skills.
Maybe they are one skill away.
Maybe they simply need to understand the gap.
That's why I wanted NextStep to do more than say:
“You're eligible.”
or
“You're not eligible.”
I wanted it to also say:
“Here's where you stand.”
and
“Here's what you could do next.”
The NextStep
The final idea is simple:
NextStep doesn't try to find every opportunity on the internet.
It tries to make the search for opportunities more meaningful for the student.
A student tells it:
Who I am.
What I know.
What I'm interested in.
What I'm looking for.
And NextStep helps answer:
What opportunities fit me?
Why do they fit me?
Where am I missing something?
And what could I work toward next?
That's the journey behind NextStep — A Student Opportunity Finder by Hari.
And this project taught me something beyond coding:
Sometimes the biggest step forward is simply knowing what your next step could be.
This website is dedicated to all my loved ones who wanted me to succeed in the future and who never wanted me to waste a moment of time on unproductive things.
Comments
Post a Comment