Job&Talent harnesses technology and AI to revolutionize job search
Madrid, 18 January 2023 - Job&Talent’s mission is to improve the working conditions and employability of millions of people internationally. We leverage cutting-edge technology to build an industry-leading platform that fuels this mission by reshaping temporary work and tearing down the barriers to job searching and hiring, as well as making the work experience more efficient and enjoyable.
While we leverage artificial intelligence to achieve this goal and make the experience for our workers smoother and easier, we always use AI responsibly, making sure it is free of biases. Our Machine Learning models do not use demographic information, but mainly historical performance on previous jobs, interest shown while browsing or applying to our vacancies offers, declared and verified experiences, as well as the distance to the job location to take the commuting time into account.
We believe in the power of AI and already leverage it to drive or support a broad spectrum of processes and features, from making our teamwork more efficient to creating a better user experience in our applications. We have been using a variety of libraries for our AI solutions alongside an AI algorithm that we developed in-house which has a pending patent. Examples of this are:
A model that is calculating how well a job fits to a worker, during and after the job is being performed. Comparing the user experience and performance with the ones of other workers on the exact same job allows us to better understand the experience and reducing the individual bias. This information is very job specific, e.g., a worker can perform the same job differently based on the clients or the work environment, this is why we calculate this fit independently for each user experience.
A model that is predicting how a new user –which has never worked with us– will enjoy a specific job. This task is very challenging since we don't have data on the previous worker's jobs, so we should estimate the likelihood of a good match by the explicit information the worker provides, as well as the implicit information we can collect. One hypothesis in this stage is that the user motivation is a good proxy for the interest in the given job.
Once we have new available vacancies, one of our goals is to notify the users that could be interested. For this specific task, we rely on different algorithms to identify the most suitable users, including the ones previously defined. In this phase, we should also calculate how many users we want to automatically notify. We have a special ML model, that we called "Invitation Logic" that needs to decide to how many people we should send these job notifications, waiting a certain time for their answer, and (based on probability) deciding if wait more, or notify new users; until the vacancies are fulfilled.