The Student Hiring Assistant team developed an AI-powered web application that guides CSU Bakersfield staff through student hiring workflows by combining a conversational chatbot with structured decision logic to help users identify the correct forms, approval paths, and grant information. The solution is built as a Next.js application with a modular Decision Engine architecture that separates intent detection, conversation state tracking, workflow rules, grant lookup, and response generation into independently testable components, and it integrates Amazon Bedrock (managed generative AI) and Amazon Bedrock Knowledge Bases (managed RAG) to answer policy questions grounded in official university documents stored in Amazon S3 (cloud storage). When a staff member asks about hiring a student, the intent detector classifies the request—workflow questions like form selection and approval routing use deterministic business rules coded from institutional policies, grant-related questions search structured spreadsheet data to retrieve Fund #, PI, analyst, project number, and CO-PI information, and open-ended policy questions retrieve supporting passages from the knowledge base and return answers with source citations. To keep the knowledge base current, the system includes an automated document ingestion pipeline using AWS Lambda (serverless compute) that monitors Amazon S3 for newly uploaded files, extracts text from PDFs and Word documents directly or from scanned images using Amazon Textract (document text extraction), normalizes the content, and triggers a Bedrock Knowledge Base sync to make the information immediately available for retrieval. The assistant follows strict safety guardrails—it never invents policies or guesses missing information, asks clarifying follow-up questions when details are incomplete, and directs users to the appropriate university office (Grants Office, Dean’s Office, Payroll) when a question falls outside its supported scope. By centralizing guidance and grounding answers in approved institutional sources, the system reduces repetitive administrative inquiries, improves consistency in form selection and routing decisions, and helps staff reach the correct next step more efficiently while keeping final decision-making authority in human hands.