Establish candidate evidence
Upload a PDF, DOCX, or TXT résumé for session-specific grounding.
Deployed public prototype · Active validation
CCSF AI Interview Coach
A document-grounded voice interview practice tool developed for CCSF Career Services in collaboration with Steve Nelson, Employment and Training Specialist for Information Technology and Computer Science.
Product case / 01
Make interview practice specific enough to be useful.
Generic question lists do not account for a student’s actual experience or the role they are pursuing. This prototype uses the résumé and job description as the working context for a realistic practice session and grounded feedback.
Problem
Students need more than canned interview questions.
A useful coach must understand what the candidate has actually done, what the target role requires, and where the interview is testing the connection between them.
Audience
CCSF students preparing for technology and computer-science interviews.
The interface reduces setup to three visible inputs: student résumé, target job description, and interview mode. An optional practice goal lets the student direct attention to a specific concern.
My contribution
Product framing, workflow design, implementation, and operating boundaries.
I translated the Career Services use case into a public prototype, structured the document-to-interview flow, designed the student-facing interaction, and made privacy and employment-decision limits visible at the point of use.
Applied AI
Context becomes preparation, not an automated judgment.
The system analyzes the supplied résumé and job description to conduct a conversational mock interview and prepare feedback grounded in the student’s materials. It is a coaching tool. It does not score employability or make employment decisions.
Workflow
Three inputs. One focused practice loop.
The design keeps the task understandable while preserving enough context to make the interview materially more relevant.
Define the target
Paste or upload the job description and select a quick, standard, or full interview.
Practice and reflect
Complete a voice mock interview and receive actionable feedback tied to the supplied context.
Human-centered controls
Privacy and purpose are part of the interface.
- Documents are described as processed in memory for the current session only.
- The tool is explicitly positioned for practice and coaching.
- The interface states that it does not make employment decisions.
- Students control the target role, interview length, and optional practice goal.
What this proves
Applied AI built around a real user and a real task.
The project connects stakeholder collaboration, product thinking, document-grounded AI, voice interaction, privacy communication, and responsible use. It shows how I bring AI into an existing service without losing sight of the person using it.