Inspiration

We, Burak and Furkan, are PhD students. We both searched for information online on which schools we should apply to and ended up choosing schools using very rudimentary estimates as simple as picking from ranking lists such as US News. We understand that careers are very hard to develop and understanding how to achieve your goals can be confounding. People can rely on online resources but there is very little to no reliability of these sources. Counselors or experts in the respective fields can help out but we have found them to be either hard to get in touch with for extended periods or are very expensive. We want to help by telling students how to actually achieve their goals by taking humans out of the loop thereby lowering costs and increasing accessibility.

On the technology side, both of us are doing PhDs in ML-coupled fields. Burak is developing methods to detect and diagnose anomalies in supercomputers while Furkan is using ML to accelerate processors and increase their performance. Between the two of us, we have ~10 papers in various layers of the computing stack and have several patents. Thus, we believe compared to our competitors, we are more capable of creating a good algorithm.

What it does

Educeron uses ML instead of humans to find data-driven scalable solutions to career counseling.

Educeron principles: -Continuity -Personalization -Democratization -Reliability -Motivation

We started by creating a prediction engine to calculate the chance of a student in getting into a grad school of their choice. We setup a website (educeron.com) to collect data and shift towards the college ecosystem over time.

How we built it

Frontend was built using ionic (v4) framework and angular js. The backend is built in django. Machine learning is used to train our model offline using python and scikit. We run the algorithm in the backend and give the user results.

Challenges we ran into

Getting the algorithm to work took a long time. There were many problems in the backend while transferring the data we had a lot to learn.

Accomplishments that we're proud of

We are able to achieve fairly reliable results (tested on our own applications and statistics).

What we learned

A fair bit about everything.

What's next for Educeron

We are thinking of continuing the development of Educeron and think that there is actually some potential to monetize in the future.

Share this project:
×

Updates