Software Development Engineer

Vinayaka Bhat

I build

Backend engineer — Java & Spring on AWS — I craft the quiet, dependable layer other systems bet on.

I turn fuzzy requirements into clean, reliable services — reaching for the right tool, not the fanciest one. Sometimes that’s a straightforward API and a well-shaped schema; sometimes it’s an event-driven, distributed design. Either way I’m big on testing, observability, and sane failure modes, and I count a quiet pager as a win. The backend is home, but I’ll climb the stack when a problem calls for it.

Vinayaka Bhat, Software Development Engineer
5+
YEARS
EXPERIENCE

Skills & Tools

Backend-focused, with data-engineering and cloud range.

How I Build

Experience

Amazon.
2021 – Present · Bengaluru, India
Software Development Engineer

~5 years at Amazon across three orgs — Kindle (reliability & infrastructure), AWS Bedrock (GenAI annotation platforms), and Cross-Border financial systems (intercompany accounting & transfer pricing) — building Java/Spring backends on AWS.

GenAI Ledger Verification Agent

Python · GenAI · Amazon Ion

Deterministic engine that recomputes accounting workbooks and self-validates journal entries against ledger records across country corridors — replacing an error-prone manual sign-off.

  • Recomputes Journal Entry Mapping workbooks in-memory and self-tests against cached formula values.
  • Validates 8 accounting segments from both batch output and integration-test logs.
  • Deterministic generator scaffolds a new corridor by editing only 2 files.

Cross-Account Transfer-Pricing Service

Java · DynamoDB · SNS · CDK

Event-driven, cross-account integration for intercompany transfer processing — with retry/idempotency and feature-flag-gated rollout.

  • DynamoDB with secondary indexes; Jackson-serialized SNS message construction.
  • Cross-account publisher spanning 6 country entities with item-level cost attribution.
  • Retry + idempotency for tax accuracy and audit traceability.

EU Cross-Border Accounting Pipeline

Java · REST · AppConfig

Multi-stage, event-driven booking pipeline delivered under regulatory constraints, rolled out gradually via config flags.

  • External REST integration for business-type classification; ledger config via AppConfig.
  • 9 production releases across 5 use cases and 10+ destination corridors.
  • Scenario-based integration testing for cross-border regulatory adherence.

Service Routing Optimization

Java · CDK · SNS/SQS/Lambda

Direct service-to-service communication layer in a high-volume, distributed transaction system.

  • Synchronous accessor libraries plus CDK-provisioned async event processing.
  • Removed 3 network hops — lower latency and cleaner failure diagnosis.

Pipeline Reliability: 64% → 91%

Java · Spring · DynamoDB

Root-caused and resolved failure modes in a high-scale, distributed workflow pipeline.

  • Fixed NPEs and batched 4,000+ record operations to eliminate timeouts.
  • Unblocked a 20-day pipeline stall from dependency ambiguity.
  • Eliminated 25 daily failures; closed 18 escalations and 5 high-severity incidents.

JDK 17 Migration

Java · Spring/RPC · Build tooling

Coordinated a runtime upgrade across a large, interconnected microservice estate with zero customer impact.

  • Migrated 17 microservices and 72 shared packages (incl. a prerequisite framework upgrade).
  • Resolved 19 compliance risks; moved onto an LTS runtime with current security patches.

Zero-Trust Authentication Migration

Java · Spring · Multi-Region

Migrated services to zero-trust auth with zero downtime via a backward-compatibility pattern.

  • 4 services across 8 regional deployments; dual-server pattern for compatibility.
  • Invented a cross-fabric proxy test method; achieved EU (DMA) regulatory compliance.

Full-Stack Platforms & GenAI Evaluation

React · Lambda · RDS · QuickSight

Built internal operations platforms and led quality for GenAI model evaluations.

  • Real-time dashboards, dispute-resolution workflow, role-based access; scaled a platform 253 → 500+ users.
  • Data pipeline scoring annotator agreement (Krippendorff’s Alpha + a custom metric) for model training.
  • Led an 8-engineer effort — 65 design docs tested, 144 CRs reviewed; automation cut manual effort ~90%.

Get in touch

Open to Senior / SDE-2 backend roles. Open to anywhere in the world.