OneApp

Overview

OneApp is a job application platform where users apply once and get matched to companies that are appropriate based on their background, interests, and experience. Our goal is to give job seekers a better, more efficient option in comparison to the status quo where people will apply to hundreds and even thousands of companies before getting an interview.

Methodology

Rather than applying to hundreds of companies each cycle, we propose creating a platform where job seekers only need to apply once and the platform will try to match each person to a role that would fit them. This way, a person can fill out their information on OneApp and be assigned an interview once the right job listing appears. Our guarantee is that each assignment will result in a face-to-face interview, but we do not guarantee that assignments will result in getting hired. We believe this platform will cut out a major obstacle and headache from the job application process.

Tech stack

For this project, we made heavy use of the Fetch.ai Agentverse platform and ASI:ONE to agentically match employers with job seekers. Each applicant's profile is processed by 6 agents hosted on Agentverse that analyze individual aspects such as past experience, education, career trajectory, and more. Afterwards, one agent advocating on behalf of the applicant and one advocating against debate whether or not the applicant should be assigned the role. Finally, all the scores and debate data are fed into Claude to make a final decision.

User Interface

Users will have access to both a web portal and a mobile application. On either portal, users will be prompted to provide identification, resume, and other information necessary to be matched to an employer. After filling out their information, users can leave the app and will occasionally be given behavioral tests or other quizzes that can boost their chances of getting an assignment.

User Data

Users will be expected to provide their name, age, resume (pdf), education, companies of interest, location, previous work experience, and expected salary. In addition, users may choose to take a behavioral exam or list up to 3 companies of interest (10 for paid users). All of this data will be kept private in a Supabase DB and will be used in consideration for job matches.

Company Interface

Employers will have access to a web console at console.oneapp.com. On this console, they can post job listings and contact prospective employees. This interface will also show key metrics such as job listing impressions and a breakdown of what types of people are being matched. After a matched company and job seeker hold an interview, we require both the company and job seeker to record the outcome of the interview (hired, rejected).

Company Data

Employers will be required to provide a detailed job description and the number of employees they want to add. We also highly encourage providing the resumes of current or past employees. Companies will have the ability to modify the relative weights of each employee metric category (ie. previous experience, behavior score, role compatibility, etc).

Algorithm Design

Our algorithm will be a multi-agent system with each agent assigning a score from 0-100 for each employee metric category. This will be performed using Fetch.ai Agentverse and the Claude API. Employees that are considered the most fit for a role will be connected with companies. In addition, we will consider interview outcomes to automatically edit the weights of each category and fine tune each agent to optimize for the highest acceptance rate post interview.

Growth Strategy

We will advertise OneApp throughout school campuses as an easier and more effective alternative to applying jobs directly through LinkedIn or company websites. With our large database of job seekers, we will work directly with employers to get job listings posted on OneApp.

Built With

• anthropic
• deepgram
• elevenlabs
• fastapi
• fetchai
• livekit
• supabase
• vercel