TECHNICAL CASE FILE / AI PRODUCT
05/2026 — PRESENT
VCareer
AI Career Development Platform
LIVE / DEVELOPMENT PENDING
A career-practice product connecting CV preparation, job criteria and live AI interviews for a VinUni student pilot. The product remains live while further development is paused pending university funding.
01 / EVIDENCE
Three signals from the real pilot.
One real-use count, one qualitative survey signal, and one expert review loop — without turning product placeholders into portfolio outcomes.
- 150+
- real student testers
- Learners in the VinUni pilot
- Survey
- higher confidence reported
- Qualitative feedback; no job-placement uplift claim
- Expert review
- CV scoring reviewed
- Product Owner, VinUni Career Services
02 / PROBLEM + SCOPE
Connect preparation to the job a student actually wants
- CV
- JD criteria
- Live AI interview
Career preparation was split between CV review and interview practice. The team began with the idea and interviews with the Product Owner from VinUni Career Services, then built one workflow spanning CV work, JD criteria and a live AI interview.
02 · Direct contribution
What I owned
TU / DIRECT
LiveKit + WebRTC baseline
Set up the baseline real-time path used by the live AI interview experience.
TU / DIRECT
CV analysis and CV-to-JD matching
Developed the deeper analysis and matching workflow that turns CV and job criteria into actionable fit feedback.
TU / DIRECT
JD Builder
Built the workflow used to create and structure job descriptions for downstream analysis and matching.
03 / DELIVERY
From discovery to a real learner pilot
06 WEEKS / DISCOVERY TO PILOT
Discovery
Idea and Product Owner interviews
The six-week team cycle began in May 2026 with problem framing and interviews with VinUni Career Services.
Build
Six weeks from idea to product
A four-person team developed the product from the initial discovery through a pilot-ready release.
Pilot
Tested by 150+ real learners
Survey feedback reported higher confidence. CV evaluation and scoring were reviewed by the Product Owner from VinUni Career Services.
AFTER THE PRODUCT WAS ALREADY LIVE
Recognition
2nd Prize · Track 4: Transform with Codex
VCareer entered the June 27, 2026 Hanoi event as an already-live product, as Track 4 required. It was not built from scratch during the hackathon and received $5,000 in OpenAI API credits.
04 / SYSTEM
Realtime in the browser, evidence in the product workflow
This section describes the wider product system. My direct ownership is listed above.
SCOPE + STATE LEGEND
- PRODUCT CONTEXT
- TU / DIRECT
- CURRENT ENVIRONMENT
- TARGET / USER CUTOVER PENDING
01PRODUCT CONTEXT
Browser + TalkingHead
TalkingHead remains part of the live interview experience; it belongs to the wider product context.
02TU / DIRECT
LiveKit + WebRTC baseline
Set up the baseline real-time path used by the live AI interview experience.
03APRODUCT CONTEXT
Next.js full-stack + managed data services
The product uses Next.js 16, Drizzle ORM, Neon PostgreSQL and managed AI services for interview, CV and matching workflows.
03BTU / DIRECT
CV/JD analysis and structuring workflows
CV analysis and CV-to-JD matching
Developed the deeper analysis and matching workflow that turns CV and job criteria into actionable fit feedback.
JD Builder
Built the workflow used to create and structure job descriptions for downstream analysis and matching.
04PRODUCT CONTEXT
Cloudflare R2 and Amazon S3 coexist
Both object-storage paths are retained in the current system; this is not a completed R2-to-S3 replacement.
TWO DEPLOYMENT BRANCHES
05ACURRENT ENVIRONMENT
Vercel / current environment
Real users currently access the production version running on Vercel.
05BTARGET / USER CUTOVER PENDING
AWS Singapore / verified target
The target architecture has been built and verified; real-user cutover remains pending.
05 / SCREENS
Six views from the live product
The screenshots show the landing flow, CV Builder, CV/job matching, live interview with TalkingHead, interview review and learner dashboard.
SCREEN 01 / 06PRODUCT CONTEXT View full screenSCREEN 01 / 06Landing pageLoading original image
Landing pageView full screen SCREEN 02 / 06TU / DIRECT-SCOPE RELATED View full screenSCREEN 02 / 06CV builderLoading original image
CV builderView full screen SCREEN 03 / 06TU / DIRECT-SCOPE RELATED View full screenSCREEN 03 / 06CV / job matchingLoading original image
CV / job matchingView full screen SCREEN 04 / 06TU / DIRECT-SCOPE RELATED View full screenSCREEN 04 / 06Real-time AI interviewLoading original imageReal-time AI interviewView full screen SCREEN 05 / 06PRODUCT CONTEXT View full screenSCREEN 05 / 06Interview review & feedbackLoading original image
Interview review & feedbackView full screen SCREEN 06 / 06PRODUCT CONTEXT View full screenSCREEN 06 / 06User dashboardLoading original image
User dashboardView full screen
06 / STATE
What exists now, and what waits for funding
PRODUCT STATE BOUNDARY
SHIPPED / CURRENT
In the current product
- Live AI interview flow with LiveKit/WebRTC and TalkingHead
- CV analysis, improvement and CV-to-JD matching
- JD Builder for structured job criteria
- Interview review and learner-facing feedback
PENDING FUNDING
Pending university funding
These items are roadmap—not shipped capabilities.
- Book mentor meetings for career advice
- Track learner progress over time
- Match candidate CVs against Career Services job postings so staff can review fit and candidates can discover suitable roles