ML Search Function – CrackedDevs

Inspiration

CrackedDevs Website missing a search function, I envisioned leveraging machine learning to transform how job seekers connect with opportunities, aiming for precision and personalization.

What it does

My ML Search Function on CrackedDevs swiftly matches users with tailored job listings. It intelligently interprets search queries, leveraging machine learning to provide relevant and personalized results.

How we built it

I created a machine learning model using Word2Vec (word embedding technique), focusing on preprocessing data (Remove HTML tags, Tokenisation, Lematization, Lowercasing, Remove urls, Replace breaking space, Etc.), developing a context-aware algorithm, and crafting a API to interact with the model.

Challenges we ran into

A significant challenge I faced was developing an effective model with a limited dataset (less than 300 job offers).

Accomplishments that we're proud of

I'm particularly proud of creating a feature that not only enhances the user experience but also showcases the practical application of machine learning in everyday solutions.

What we learned

Apply word embedding with a small dataset

What's next for ML Search Function – CrackedDevs

• Extend the application of machine learning models to include additional functionalities, such as enabling tag-based searches.

• Improve the model's performance by conducting further training with a larger and more diverse dataset.

• Incorporate the model into the existing CrackedDevs API.

Built With

• api
• machine-learning
• numpy
• panda
• python