🌱 Java AI Engineer | 2.8+ Years Experience
🤖 Architecting Agentic AI Systems & Distributed Backend Platforms
🏗️ Focus: High-Concurrency, Event-Driven Microservices & Java 21 Performance
📊 Obsessed with: System Design, Observability (Prometheus/Grafana), & Thread Safety
⚡ Mission: Building production-grade software that survives the "Flash Sale" load.
📧 Email: mdnazir2608@gmail.com
🌐 Portfolio: nazir2608.github.io/portfolio
A selection of my most complex architectural work.
| Project | Deep Technical Focus | Key Tech |
|---|---|---|
| Online Auction System | Anti-snipe logic, concurrent bid resolution, & real-time event processing. | Java 21, Redis, Kafka |
| Vision-AI Chatbot | VLM (Vision-Language Models) for analyzing diagrams & handwriting via Spring AI. | Ollama, RAG, VLM |
| Ecommerce Microservices | Full observability suite, distributed tracing, and circuit breaker patterns. | Spring Cloud, Grafana |
| High-Perf URL Shortener | Tier-based rate limiting, geo-analytics, and low-latency redirection. | Redis, Java 21, SQL |
- High-Throughput: Experienced in handling race conditions (Redis locking/distributed locks).
- Event-Driven: Decoupling services using Apache Kafka and Domain Events.
- Modern Java: Leveraging Java 21 Virtual Threads for high-concurrency scaling.
- Reliability: Implementing Circuit Breakers (Resilience4j) and Rate Limiting (Bucket4J).
- Agentic Workflows: Multi-step AI task orchestration with state management.
- RAG: Advanced Retrieval-Augmented Generation for private data grounding.
- VLM: Integration of Vision-Language Models for multi-modal backend processing.
- Currently Mastering: Graph Modeling with Neo4j for recommendation engines.
- Exploring: Multi-tenant SaaS isolation strategies at the database level.
- Recent Achievement: Built an AI-powered invoice analyzer with 95% OCR accuracy.
💡 Code that works is the baseline.
🚀 Code that scales is the requirement.
🔥 Code that survives production is the art.
⭐ Building reliable, scalable systems — one high-concurrency challenge at a time.