Ai.resumeshortlister

• 

  1

• 

  2

\section*{Inspiration} The idea for \textbf{AutoShortlist.AI} came from observing how much time recruiters spend manually checking resumes. During one placement drive, a recruiter shared that they went through over 1,000 resumes in just two days --- and still missed a perfect candidate. That moment made us ask, \textit{"What if resumes could shortlist themselves?"}
We wanted to build something that could save time, remove bias, and make hiring fair for everyone. That’s how AutoShortlist.AI was born --- an intelligent system that automates resume screening using AWS-powered AI agents.

\section*{What it does} \textbf{AutoShortlist.AI} is an autonomous resume shortlisting system built on AWS. It automatically reads, analyzes, and scores resumes based on job descriptions using AI. The system then provides recruiters with a ranked shortlist --- saving up to 80\% of review time while ensuring fair, unbiased evaluation.

\section*{How we built it} We built AutoShortlist.AI using \textbf{Python} and \textbf{AWS Agentic AI services}:
\begin{itemize} \item Amazon S3 for secure file storage. \item AWS Textract to extract text from resumes. \item Amazon Bedrock for AI-based skill and job matching. \item AWS Lambda to automate workflows. \item Amazon SageMaker for continuous learning from recruiter feedback. \item Amazon DynamoDB to store shortlisted results. \end{itemize}

We used a multi-agent design, where four specialized agents --- \textit{Parser, Analyzer, Feedback, and Reporting} --- work together seamlessly.

\section*{Challenges we ran into} \begin{itemize} \item Handling different resume formats (PDF, DOCX, scanned files). \item Designing an unbiased scoring system using LLMs. \item Ensuring smooth data flow between multiple AWS services. \item Managing security and permissions for sensitive data. \item Tuning Bedrock prompts for more accurate candidate-job matching. \end{itemize}

\section*{Accomplishments that we're proud of} \begin{itemize} \item Built a fully automated shortlisting system with 80\% time savings. \item Achieved 95\% text extraction accuracy using Textract. \item Created a scalable, serverless workflow using AWS Lambda. \item Designed a feedback-driven learning loop with SageMaker. \item Developed an AI tool that promotes fair, bias-free hiring. \end{itemize}

\section*{What we learned} \begin{itemize} \item How to integrate multiple AWS services into one intelligent system. \item The power of multi-agent architectures for autonomous operations. \item The importance of data fairness and feedback loops in AI recruitment. \item How small design decisions can greatly impact accuracy and scalability. \end{itemize}

\section*{What's next} \begin{itemize} \item Integration with LinkedIn and other professional networks. \item Adding voice and video screening using Amazon Transcribe and Comprehend. \item Expanding to multi-language support for global hiring. \item Building a recruiter dashboard for real-time insights and analytics. \item Preparing for enterprise-level deployment and testing with real HR data. \end{itemize}

Built With

• awsbedrock
• awsdynamodb
• awslambda
• python