Inspiration
Our team's collective experience revealed the challenges of securing employment or internships. The laborious task of sifting through job descriptions and ensuring their legitimacy proved a significant time drain, highlighting the need for a more efficient solution.
What it does
Scan Unlikely is an integrated solution (Chrome Extension and Website) crafted to enhance job seekers' experience by simplifying their search journey and protecting them from potential scams. Our extension analyzes job offers, generating a concise report including job details and the likelihood of it being fraudulent, leveraging a pre-trained LLM. Users can conveniently save job opportunities they encounter through our extension for future reference.
How we built it
Scam Unlikely is a multifaceted project encompassing a Website, a Chrome Extension, and Backend infrastructure.
Backend Infrastructure: We leveraged FLASK and PostgreSQL to ensure a robust foundation for data management, enabling seamless integration and scalability.
Integration of OpenAI API: ChatGPT was integrated to generate concise job description summaries and conduct associated risk assessments for job offerings.
Clerk for User Authentication:Clerk was implemented for secure user authentication, providing a seamless and secure login experience.
Website Development: The Website's graphical user interface (GUI) was crafted with React, ensuring a seamless and responsive user experience.
Extension Interface Engineering: The Extension's interface was designed to help users to parse job postings directly from their browser, generating a job report and saving it for future reference.
Hosting: Hosting everything on AWS ensures scalability and reliability.
Challenges we ran into
We had only 24 hours to build the entire system, including two GUIs (website and extension) and a backend. We mainly struggled with making the extension run as expected.
Accomplishments that we're proud of
Brainstorming a whole system architecture, developing it, and deploying it in 24 hours is. We used a variety of technologies to make the system run effectively.
What we learned
Prioritizing Features implementation. Making a plan B.
What's next for Scam Unlikely
Regarding scalability, Scam Unlikely seamlessly integrates with Indeed and LinkedIn, employing customized parsing methods for their DOM structures. Our future efforts will prioritize extending parsing capabilities to diverse job search platforms. We also aim to introduce job matching functionalities and improve user account integration.
Moreover, we plan to incorporate a user feedback feature to refine our risk assessment process. Additionally, we intend to develop and train a proprietary model capable of leveraging task-specific GET data, offering more personalized outcomes compared to pre-trained LLM models such as GPT."
Log in or sign up for Devpost to join the conversation.