Building a Chatbot for Interview Preparation using NLP

Challenge Background

The challenges candidates face in the traditional interview process, particularly for roles requiring a range of skills. It emphasizes the lack of comprehensive evaluation of traits like behavior, communication, empathy, leadership, and problem-solving. Additionally, the absence of tailored resources and realistic mock interviews leaves candidates unprepared for these crucial assessments. The HR Interview Prep Chatbot project aims to address these issues by offering a personalized, immersive preparation experience, ultimately empowering candidates to approach interviews with confidence and competence.

The Problem

Omdena Hyderabad Local Chapter is working on a LLM based chatbot that can conduct mock interviews for aspiring candidates, evaluate the responses of job aspirants during mock HR interviews and provide quantitative, personalized feedback for each candidate along with recommendations to develop their skills to successfully navigate HR rounds of interviews.

This project will focus on those roles or positions which involve a thorough assessment by HR professionals to understand the behavioral traits, communication skills, empathy, leadership, and problem solving skills applicable for the role. Some examples of roles that will come under this project's purview are teaching positions, managerial or leadership roles and sales related positions. The project will focus on the top 3 roles for which candidates can benefit from the mock interview process. The top 3 roles will be identified at the end of the data collection phase. 

The data collection phase will focus on collecting a varied set of interview questions asked during HR rounds and the valid answers (and inappropriate answers as well) for different circumstances dependent on the job description, domain requirements and candidate experience. This will be done for the top 10 roles that have the most number of job postings in various online forums such as Indeed, Linked In, Glassdoor etc. The collected data will be reviewed to identify the three most suitable roles for further implementation.

A domain exploration group will work with qualified HR professionals and soft skills trainers (or linguists) to further review the collected data and provide inputs on the different aspects that need to be evaluated in addition to the veracity and suitability of answers provided by the candidates.

Based on the above inputs, the pre-processing phase will focus on building the application and the pipeline that interacts with end users, conducts mock interviews, collects responses, pre-processes the inputs from end users, invokes the evaluation model(s) and presents the results and recommendations back to end users.

A modeling team will identify the suitable models, LLMs, decide on the approach for working with LLMs (fine tuning or retrieval-augmented generation - RAG) and design the pipeline for doing evaluation, and identifying recommendations.

Goal of the Project

The objectives of this challenge are: 

• Build a platform to help job aspirants to attend mock interviews and identify areas of improvement
• Explore the possibilities of LLM based apps in human development
• Build skill sets of participants in NLP, LLM based apps, Prompt Engineering and Retrieval Augmented Generation (RAG)

Project Timeline

What you'll learn

1. Large Language Models Retrieval Augmented Generation (RAG)
2. Prompt Engineering Chatbots
3. Audio Signal Processing (Speech to Text and Text to Speech)
4. Natural Language Processing Model Deployment