Using ChatGPT-like models in the talent hiring and screening

Hello

can drastically improve candidate screening turnaround times and improve quality of hires.

Image by Clem Onojeghuo from Unsplash

Impacted Areas

Industry

All industries

Company Size

Most large organisations

Time to Value

POC in 8-12 weeks

Problem Space

Predictive analytics

Large Language Models

Automation

Machine Learning

Technology Stack

ChatGPT-like LLMs

Python

ScikitLearn

Kubernetes & Kubeflow

Business Context

The hiring process is a time-consuming process that often cannot guarantee effective results. This means that the ability of an organisation to provide the right skills for the job at hand is based on a series of often manually conducted steps driven by the organisation’s Human Resources (HR) department. The steps become inefficient as the volume and nature of the jobs change over time.

Businesses are looking for ways to improve the speed and effectiveness of identifying, screening and pre-processing candidates. The goal is to provide the business with a shortlist of high quality candidates that have already been matched to the job and skillset requirements, organisational culture, candidate career growth alignment and salary expectations.

The use case we explore specifically focuses on impacting the following hiring steps:

  • Job profiling and specification
  • Candidate screening
  • Pre-screening activities such as a qualification call or video interview
  • Shortlisting of candidates for in-person interview


Challenges

High volume of applicants

Large organisations often receive a large number of job applications, making it difficult to screen and select the best candidates efficiently.

Time-consuming process

The hiring process can be lengthy and complex leading to delays in filling open positions, potentially with candidates that are not a good fit for the position.

Inconsistent and biased approach to screening candidates

Ensuring that the hiring process is consistent and fair for all candidates can be a challenge. With multiple people involved in the process, there is a risk of bias or inconsistencies in the evaluation of candidates.


Solution

A continuously learning recruitment advisor

The solution will incorporate a feedback loop from existing internal and available external data to generate baseline job profiles for the organisation. These profiles are based on existing positions that have been through the recruitment value chain.

The data is fed as inputs into a machine learning model including the use of a ChatGPT-like Large Language Model (LLM) to generate the profiles based on pre-trained and additional training on the model. The output is a job profile structured with the core requirements for the position.

What makes this solution unique is how as data collection improves, the entire value chain benefits.

Recruitment screening using a context-aware LLM for enhanced question/response interaction and probing

The actual recruitment screening process will utilise the LLM to improve screening questions asked each candidate, with additional probing follow-up questions in order to surface more relevant insights for the screening and matching process.

Machine learning classification models can be developed to predict the fit of the candidate to the organisation and position advertised. This will take in more advanced features from the AI-driven screening process and train against the success placement of candidates within the organisation.

At Melio we apply our proven methodology FLUID4ML to ensure that the solutions we build are focused on solving business problems that are iteratively developed to release value early. We build with durability and long-term thinking in mind to ensure that our solutions are sustainable and extendable.


An AI-driven recruitment screening advisor will help HR teams and organisations unlock value rapidly:

🕑

Improved candidate selection and screening turnaround times

Improved candidate quality and fit-for-role

💰

Reduction in hiring costs

🟰

Consistency in screening and validation process

📜

Improved job profile definitions


Contact us for more information on how we can assist you.