Software Engineer with 4+ years of experience building scalable backend systems,
distributed services, and cloud-native applications.
Currently working at Serko building platform services and AI-powered developer
tools, with hands-on experience in:
- Java & Spring Boot
- Microservices & Distributed Systems
- AWS & Azure
- Event-Driven Architecture
- RAG & LLM Applications
- Model Context Protocol (MCP)Languages
Backend & Frameworks
Cloud, DevOps & Tooling
Applied AI & LLMs
🔔 Notification Platform - Event-Driven Multi-Channel Delivery System
A production-grade, event-driven notification platform supporting email, SMS, push, and webhook delivery channels, built on a hexagonal architecture for clean separation between domain logic and infrastructure.
| Aspect | Details |
|---|---|
| Stack | Java 23, Spring Boot, Apache Kafka, Redis, PostgreSQL, Docker Compose |
| Scale | 5-topic Kafka design for multi-channel event routing and delivery orchestration |
| Performance | Asynchronous, non-blocking event consumption with channel-level isolation |
| Security | Domain-isolated hexagonal boundaries; provider credentials externalized |
| Impact | Reusable, provider-agnostic notification core adaptable to any messaging backend |
| Repository | github.com/abhim8/notification-platform |
Designed around ports-and-adapters principles, the platform decouples core notification logic from third-party providers (SendGrid, Twilio, FCM), enabling clean testability and provider interchangeability without touching business logic. Kafka topics are structured to support retries, dead-letter handling, and channel-specific throughput tuning, with Redis used for idempotency and delivery-state caching.
🧠 AI Orchestration Platform - Multi-Agent Planning & Execution Engine
An AI orchestration platform built with Spring AI and MCP, capable of generating, validating, and executing multi-step AI-driven execution plans with dependency-aware parallelism.
| Aspect | Details |
|---|---|
| Stack | Java, Spring Boot, Spring AI, Gemini, Model Context Protocol (MCP), Multi-Module Maven |
| Scale | Multi-module architecture isolating planning, tooling, and execution concerns |
| Performance | Parallel execution derived purely from step-level dependsOn graph resolution |
| Security | Isolated planner/ sub-package boundary around all Spring AI integration surfaces |
| Impact | Reusable execution-plan validator with cycle detection & confidence thresholding |
| Repository | github.com/abhim8/ai-orchestration-platform |
The platform wraps generated plans in a PlanGenerationResult envelope around a structured ExecutionPlan, with an ExecutionPlanValidator performing cycle detection and confidence-based gating before execution. Partial failures are handled gracefully with an inline executionTrace, and custom tools (e.g. ResolveRelativeDateTool) are exposed to the model via Spring AI's @Tool annotation - keeping all LLM-specific logic cleanly isolated from the rest of the system.
- Software Engineer 2, Serko
Jul 2025 — Present - Senior Software Engineer, LeadSquared
Apr 2025 – Jul 2025 - Software Engineer, LeadSquared
Apr 2024 – Mar 2025 - Associate Software Engineer, LeadSquared
Aug 2023 – Mar 2024 - SDE Intern, LeadSquared
Jul 2022 – Jul 2023
| Degree | Institution | Year |
|---|---|---|
| B.Tech, Computer Science & Engineering | CMR Technical Campus, Hyderabad | 2023 |
- 💬 Happy to discuss Java, Spring Boot, distributed systems, and backend architecture
- 🧠 Passionate about scalable and reliable production systems
- 🏍️ Motorcycle enthusiast & adventure seeker - always up for a road trip!
- 🎮 Enjoy hackathons and building practical side projects
- ⚡ Fun fact: I believe there are only two ways to write error-free programs; only the third one works 😄