I am a Backend Engineer specializing in Java, Spring Boot, and distributed systems. I design microservices built to remain correct under retries, concurrency, and partial failure—focusing heavily on idempotency, event-driven messaging, and resilient integrations.
- 🔭 Currently a Software Developer at Amdocs, owning production microservices and driving batch ETL performance optimization for T-Mobile.
- ⚡ Core Impact: Boosted transaction reliability by 20%, optimized Redis session token-storage by 35%, and scaled concurrent transfers to eliminate lost updates.
- 💬 Ask me about: Defending against double-spends, transactional outbox patterns, or tuning SQL batch windows.
Java 21 • Spring Boot 3 • JWT • JPA • Event-Driven A digital wallet backend built to handle critical payment system failure modes:
- Concurrency Control: Implemented optimistic locking (
@Version) with bounded retry loops executing outside the Spring transaction proxy to prevent lost updates under low contention. - Double-Spend Prevention: Enforced strict idempotency keys driven by database unique constraints to safely stop concurrent duplicate requests.
- Reliable Messaging: Leveraged a Transactional Outbox pattern paired with at-least-once relay and consumer deduplication to guarantee eventually consistent processing.
Spring Batch • SQL Tuning • Table Partitioning • Amdocs (T-Mobile)
- Problem: Peak run times were blowing past the designated batch windows due to read-heavy bottlenecks.
- Solution: Profiled operations and introduced SQL tuning alongside database table partitioning to enable independent slices to run in parallel steps without write contention. Cut peak runtimes significantly.
- Portfolio & Deep Dives: patil2001.github.io/portfolio
- LinkedIn: linkedin.com/in/buddhapp
- Email: patilbuddhaprakash@gmail.com