Rizz-Umé

Rizz-Umé is a Telegram bot that helps students streamline their job application process by providing personalized feedback on their resumes, job recommendations and even writing cover letter drafts!

Comment

Inspiration

As students, creating our resume may seem like an extremely daunting task with numerous factors that we have to take note of such as formatting, word structure and vocabulary. It quickly becomes overwhelming, especially without someone to guide you or give you feedback to your specific situation. This pressure is further compounded due to the importance of trying to secure that summer, winter or semester internship for a university student.

As students ourselves in similar situations, struggling to craft the perfect resume to apply for our internships. We wanted to create a way for us to get feedback on our resumes as and when we need in addition to the help that our universities may already provide us.

Introducing Rizz - Umé, a personal career consultant telegram bot!

Using Rizz - Umé, we aim to help students streamline their job application process by providing them the resource and support to improve their resumé writing abilities helping to tailor it to specific job descriptions. Additionally, Rizz - Umé also helps you by recommending jobs based on your resume.

With the help of Rizz - Umé we aim to remove the guesswork of creating one’s resumé, helping students focus on the other aspect of the job application process such as the interview or the online assessments that the job may present to us.

Why a telegram bot?

We decided to leverage telegram as the platform to host our application as Telegram has a large user base with numerous active users, many of which are students in Singapore which is our targeted demographic. This helps with the convenience of using our bot and prevents the inertia of having to download a new application just for this single function.

How we built it

Leveraging the Telegram Bot API as our frontend solution, we were able to deploy a lightweight and efficient text user interface that our users can upload their resumés and get feedback from the bot.

For our backend solutions, we are using Node.js with the associated modules to handle reading in the PDF files and validating it before passing it into the Open.AI API. Which reads the resumé giving its response which we then tailor to the user. Additionally we used Firebase to store the resumé, so that users do not have to reupload their resumes every time they start the bot.

Finally we are using the recently released Open.AI API to handle our Natural Language Processing (NLP) as well as reviewing and producing feedback of the resumés.

Challenges we ran into

Starting off, we had trouble creating the application on our desired platform, as we had little to no experience coding a Telegram bot. In addition we had issues when trying to make the conversion with the bot seem natural; trying to make the bot respond in a human-like manner.

Next we had trouble utilising the Open.AI API, due to how the AI was trained, we found it difficult to find the right way to word our prompts to get a desirable output. Finally, with how new Open.AI is, we had a problem with the stability of the platform, as it would sometimes go down due to large amounts of traffic causing our bot to return a null output when a resume request is sent in.

Accomplishments that we're proud of

However, our challenges are also what we are most proud of for this hackathon. We are proud to have created a valuable tool for students that streamlines the job application process and provides personalised feedback on their resumes in just 48 hours as well as how we were able to utilise the cutting edge and latest technology in artificial intelligence to produce an application that we would use ourselves even after the hackathon is over. We are proud to have seamlessly integrated a robust and reliable database within a tight deadline, allowing for efficient storage and retrieval of resumes for our users.

What we learned

We learned that maintaining good coding practices is imperative to the success of application development. Even though this application is relatively small scale in size, maintaining a Model, View and Controller (MVC) framework helps to keep all our files and code in order so that we did not waste time digging around directories for the files we needed. We also learned how to make calls to an AI API to make full use of its capabilities for our application, integrating it with a telegram bot.

Finally, we were exposed to conducting market research in trying to decide what platform we should use to host our application.

What's next for Rizz-Umé

In addition to providing personalised feedback on resumes, Rizz - Umé also could be expanded to have the capability of recommending real-time job listings to users based on the information in their resume. As we continue to improve and update our AI model, the feedback provided to users will also improve in terms of accuracy and relevance. Furthermore, we are exploring the possibility of incorporating our own machine learning model to convert resumes in non-text format to text, further enhancing the user experience and compatibility of our bot.

Try it out yourself!

https://github.com/kingsmil/techfestival

  • Set up a firebase real time database, get the name the admin sdk key json file private.json, add it to the local repo
  • Create a bot using @Botfather and get the API
  • Obtain OpenAI key
  • configure .env file using example given
  • run yarn install in your terminal
  • run node index.js

Usage

After you /start the bot on telegram, here's what you are able to do using the commands in the keyboard provided.

  • Upload Resume
  • Improve Resume
  • Match Resume to Job Description
  • Draft Cover Letter to Job Description
  • Job recommendation based on Resume
  • Ask any custom question and add resume to the question if so desired
Share this project:

Updates