Case index
Case File 005Work · Hyundai AutoeverStatus: Active2025

AI Vehicle Evaluation — Real-Time PDF Generation & Token Auth Engine

Real-time PDF generation from vehicle evaluation data processing ~10k evaluations/month with caching and token auto-refresh.

Monthly throughput
~10K / Month

Vehicle evaluation PDFs generated reliably.

Auth resilience
Auto-Refresh

Zero downtime token & refresh-token renewal.

Performance
API Data Caching

Optimized PDF generation using stored payloads.

Opening brief

Vehicle evaluation reports generated from third-party vendor APIs suffered latency bottlenecks and failed when authorization tokens expired during active evaluations.

Outcome on record

Engineered a real-time PDF generation system processing ~10k evaluations/month, optimized PDF rendering with cached API data, and automated token refresh handling for zero-downtime vendor access.

Exhibit A

Project description

Filed statement

Engineered a real-time PDF generation pipeline from complex vehicle evaluation datasets processing ~10,000 evaluations per month.

Optimized PDF creation and rendering times by caching and storing vendor API payload data directly in the database.

Built an automated token and refresh-token authentication handler that detects expired tokens and renews them transparently before API requests execute.

Ensured high availability and uninterrupted access to vehicle evaluation records for inspectors, buyers, and dealers.

Exhibit B

Why I built this

Filed statement

Vehicle evaluation inspectors and dealers needed instantaneous, reliable PDF generation during vehicle assessments.

Third-party vendor API tokens frequently expired during active evaluations, causing evaluation failures.

Exhibit C

Process and execution

Approach

How the work moved

  • Built a real-time PDF generation pipeline from complex vehicle evaluation datasets.
  • Optimized PDF compilation speed by persisting and caching vendor API payload data in the database.
  • Engineered automated token and refresh-token lifecycle management that renews expired credentials on the fly before API requests execute.
  • Scaled the evaluation generation pipeline to comfortably process ~10,000 evaluations per month.
Responsibilities

What I owned

  • Led the work as Associate Consultant - Java Backend Developer.
  • Turned messy context into a clearer operating model, interaction direction, and rollout shape.
  • Worked across stakeholders to move the project from framing into execution.
Exhibit D

Notable decisions

Decision

Automated token renewal interceptor

Intercepted vendor API requests to verify token validity, automatically renewing expired tokens to ensure uninterrupted data access.

Decision

Cached API data payload rendering

Stored raw API evaluation payloads locally to eliminate repeated vendor network calls during PDF compilation.

Exhibit E

What I learned from this

Filed statement

Proactive token renewal interceptors eliminate intermittent auth failures when integrating third-party vendor APIs.

Payload caching in local database storage slashes PDF compilation latencies.