Backend Software Engineer. 5+ years owning high-throughput distributed systems in payments & SaaS. Hands-on with React when the UI is the bottleneck.
Note: this portfolio ships live edge endpoints — /api/health and /api/stats. Open them.
Who runs this service. Bio, education, signals, and the engineering principle I keep coming back to.
I'm a Software Engineer with 5+ years building and scaling high-throughput distributed systems in payments, SaaS, and retail-tech.
At CP Axtra, I'm building backend microservices for a payment-related invoice processing system — reactive WebFlux, Kafka event pipelines, Keycloak OAuth2, distributed ShedLock schedulers, and a Testcontainers integration framework targeting 75% instruction coverage.
Previously at AppDirect I designed and owned high-availability microservices for Billing, Checkout, Notifications, and Payments — supporting thousands of concurrent requests with sub-200ms p99 latency. I led the migration from RabbitMQ to Kafka, achieving a 4× throughput improvement for billing workflows processing millions of events per day.
My core expertise spans microservices architecture, event-driven systems, JVM performance tuning, and building production-grade observability stacks with Prometheus and Datadog APM.
B.Tech + M.Tech · Electronics & Communication
2015 – 2020
Great systems aren't just fast — they're observable, resilient, and maintainable. I build for the 3am incident, not just the happy path.
— principle
What this service runs on. Primary tags marked with a leading dot.
Where this engineer has shipped. Newest deploy first. Click a row to expand the changelog.
Production-grade systems built with performance and reliability as first principles. Each card is a service this engineer designed and shipped end-to-end.
Java 21 + Spring Boot router with regex routing, three balancing strategies, hand-rolled circuit breaker, token-bucket rate limiter, and live WebSocket telemetry. Validated at 10K+ RPS, p99 < 20 ms, ~210 MB resident.
Production-shaped real-time chat with JWT/RBAC, S3-backed file sharing with antivirus scanning, transactional Kafka outbox for audit, and per-conversation ordering. Sync delivery and async fan-out fail independently.
Layered Spring Boot service for users, catalog, carts, and orders with Stripe payments, robust validation, standardised error handling, and full Swagger docs. Pragmatic foundation for an e-commerce MVP.
Production-grade REST API for tracking job applications, interviews, notes, and reminders. Spring Boot 4, Java 25, Spring Modulith, Redis caching, Flyway migrations (41 versions), AI CV tailoring via Groq, WebAuthn passkeys, TOTP 2FA, Pro tier billing with Lemon Squeezy webhooks, and full OpenAPI codegen for the companion Next.js frontend.
Full-stack companion frontend for the interview-tracker API. App Router with server-side prefetch + HydrationBoundary, drag-and-drop Kanban, recharts analytics, AI JD analysis, cover-letter generation, CSV/JSON bulk import, PWA install prompt, TOTP + passkey login flows, Playwright smoke tests, and a Pro paywall wired to the 402-response pattern.
Spring Boot reactive backend for an LLM-powered knowledge assistant: retrieves and reasons over indexed engineering docs, code repos, APIs, and production systems. Supports semantic + hybrid search, multi-provider LLM routing (Anthropic, OpenAI, Ollama), ingestion pipelines (GitHub, GitLab, Confluence, web), Neo4j knowledge graphs, and JWT+OAuth2 auth with email-password fallback.
Full Next.js 16 App Router frontend for the EKA knowledge assistant. Features LLM chat with SSE streaming, hybrid code search, ingestion source management, interactive Neo4j knowledge graph, usage analytics, observability dashboards, admin user management, and email+OAuth2 auth. Radix UI primitives, TanStack Query server state, Zustand client state, Tailwind v4 theming.
Open-source data sync platform — CDC, snapshots, pipeline designer, multi-DB connectors
Java 25 + Spring Boot control plane for moving data between databases. Hexagonal (ports & adapters) architecture around a `syncflow-core` domain: Debezium CDC capture, snapshot backfill, a versioned pipeline designer with conflict detection and a transformation rule engine, a workflow scheduler, and a distributed agent fleet. Inbound adapters in `syncflow-api`, and outbound connectors behind a plugin SPI (`CdcProvider`, `SnapshotProvider`, `DestinationWriterProvider`). Multi-tenant data isolation, Kafka transport with a replayable DLQ, and a Micrometer → Prometheus/Grafana observability stack.
Contributions to the broader ecosystem — PRs merged, issues filed, and projects maintained outside my own repos.
Open to senior backend roles, system design discussions, or interesting technical collaborations. The on-call engineer is currently me.
Currently serving notice. Looking for Senior / Staff backend roles — Java, Spring Boot, Kafka, React, distributed systems. Open to remote or Pune / Bengaluru.
Pulled live from GitHub and LeetCode by an edge route handler this portfolio ships. Cached at the edge for 6 hours; falls back gracefully on rate-limit.
System diagrams for the services above. Switch modes to trace a request, or to see what happens when things break.
API gateway · circuit breaker · dynamic LB · 10K+ RPS
Five-tier architecture. Single ingress, regex-matched routing, hand-rolled circuit breaker per route, 3 pluggable LB strategies, and a backend pool of stateless replicas. Observability via Spring Actuator + Prometheus + a WS dashboard for live tail.
tip: hover any node for details · use Sequence to trace a single request · Failure mode shows what gives way first.