Varun
Kumar K

Software Engineer I

1. 01Led the migration of a large monolith handling text, email, and fax channels into a channel-specific distributed microservice architecture (Java, Spring Boot) — cutting latency by 60% and improving maintainability and monitoring.
2. 02Own a distributed, microservice-based targeting platform evaluating 100M+ patients daily against eligibility rules and publishing decisions to downstream systems.
3. 03Delivered 10+ revenue-generating programs on the targeting platform, taking each from requirements through production rollout.
4. 04Engineered an AI-powered classification system interpreting user responses — including adverse events and stop requests — at 99% accuracy, feeding results into downstream automation workflows.
5. 05Implemented PostgreSQL partitioning on AWS RDS to scale database workloads under growing traffic; led a PostgreSQL version upgrade and a Terraform-based database migration, driving 40% higher throughput.
6. 06Built an automated testing framework (TestNG) that eliminated manual regression testing and lifted release success to 95%; raised unit test coverage from 30% to 90%+ and added SonarQube quality gates.
7. 07Instrumented services with Prometheus and Grafana for real-time pattern detection across messaging traffic, adding alerting that cut production monitoring effort from hours to minutes.

Software Engineer — Intern

1. 01Built configuration-management REST APIs that let teams update service configurations through endpoints instead of manual database queries, removing the need for direct production database access.

Multi-Channel Messaging Gateway

A distributed messaging platform fronted by a centralized Spring gateway that exposes secure, rate-limited REST APIs for SMS, WhatsApp, email, and fax — each channel backed by its own independently deployable microservice. A self-hosted take on the vendor-abstraction platforms I build in production.

1. 01Centralized Spring gateway exposing secure, rate-limited REST APIs, with ingest, routing, and per-channel delivery split into services that scale and deploy independently.
2. 02Routing layer dispatches each message to the single channel named on the request; adding a provider is just a new adapter (Strategy pattern), with no change to core routing.
3. 03Channel-specific delivery workflows with decoupled provider integrations, so a vendor swap never reaches the caller.
4. 04Kafka pipeline using the transactional-outbox pattern, with retry and failure handling plus duplicate suppression that guarantees at-most-once user-visible delivery.
5. 05Exponential-backoff retries with a dead-letter queue, per-vendor rate limiting (Redis), and automatic multi-provider failover.
6. 06Prometheus + Grafana dashboards for per-channel latency and delivery success, runnable locally via Docker Compose.