Abhishek Kulkarni’s Post

I got tired of catching the same Java mistakes in code reviews. String concatenation in loops. Mutable state leaking through getters. What was missing was a consistent layer that actually understands context. So I built one. It's called the 𝗝𝗮𝘃𝗮 𝗕𝗲𝘀𝘁 𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗲 𝗖𝗵𝗲𝗰𝗸𝗲𝗿. It's an agentic skill that plugs into your editor, and from that point on your AI agent knows Java best practices from modern Java 21–24 features down to JVM internals, collections, thread safety, and design patterns. It works across 𝙑𝙎 𝘾𝙤𝙙𝙚, 𝘾𝙪𝙧𝙨𝙤𝙧, 𝙒𝙞𝙣𝙙𝙨𝙪𝙧𝙛, and any other agentic editor. The idea behind agentic skills is simple but underappreciated. Instead of relying on your AI to guess what good code looks like, you give it a precise, versioned definition of your standards. It sits at the center of your development environment and applies that knowledge consistently across every file, every session, every teammate. For teams this is meaningful. Your best engineer's instincts stop being a bottleneck and start being infrastructure. The skill is live at https://lnkd.in/e5gnAW-v feel free to try it. Would love to hear what you think. #Java #SoftwareEngineering #AI #DeveloperTools #OpenSource

⏳ Day 21 – 1 Minute Java Clarity – Polymorphism Explained Simply 🎭 Same action… different behaviour! 🔥 📌 What is Polymorphism? 👉 “Poly” = Many 👉 “Morphism” = Forms ✔ One method behaves differently based on the object 👉 It is the core of OOP and makes code flexible & reusable 📌 Types of Polymorphism 1️⃣ Compile-Time Polymorphism 👉 Achieved using Method Overloading class Calculator { int add(int a, int b) { return a + b; } double add(double a, double b) { return a + b; } } ✔ Same method → different parameters ✔ Decided at compile time 2️⃣ Runtime Polymorphism 👉 Achieved using Method Overriding class Animal { void sound() { System.out.println("Some sound"); } } class Dog extends Animal { void sound() { System.out.println("Dog barks!"); } } Animal a = new Dog(); a.sound(); // Dog barks! ✔ Parent reference → child object ✔ Decided at runtime 💡 Real-time Example 🎮 Game Character Same action → "attack()" Warrior → uses sword ⚔️ Archer → uses bow 🏹 Mage → casts spell 🔮 👉 Same method → different behaviour ⚠️ Interview Trap 👉 Can we achieve polymorphism without inheritance? ✔ YES (Using Method Overloading) ❌ NO (for Runtime polymorphism – needs inheritance) 💡 Quick Summary TypeMethodTimeExampleCompile-TimeOverloadingCompile Timeadd()RuntimeOverridingRuntimesound() 🔹 Why Polymorphism matters? ✔ Cleaner code ✔ Easy to extend ✔ Reduces duplication ✔ Helps in real-world system design 🔹 Next Topic → Encapsulation in Java 🔐 Which type do you find tricky — Compile-time or Runtime? Drop 👇 #Java #JavaProgramming #Polymorphism #OOPs #CoreJava #JavaDeveloper #BackendDeveloper #Coding #Programming #SoftwareEngineering #LearningInPublic #100DaysOfCode #ProgrammingTips #1MinuteJavaClarity

Most AI agents have no guardrails. This Java one has four. All of them were tested by breaking it. AgentOrchestrator.java is done. ReAct loop — Thought → Action → Observation → repeat — running in production Java. Not a notebook. Not a demo. A Spring Boot endpoint that returns { answer, toolCallsUsed, iterationsUsed, costTokens }. Here’s what I baked in before anything else: 10 iterations default. 20 hard cap. Full stop. No config flag that lets some engineer bump it to 1000 at 2am “just to test.” Two tools only. DBLookupTool hits PostgreSQL. HttpCallTool hits external APIs. ToolRegistry.java manages registration. Clean boundary. Easy to audit. The guardrail work was where I stopped feeling clever and started feeling scared. I intentionally broke everything: — Infinite loops. The cap caught it. — Hallucinated tool arguments. The type validation caught it. — Context window overflow. The ceiling caught it. None of these edge cases are in the Python tutorials. Because Python tutorials assume you’re demoing, not deploying. The asymmetry I’m playing is simple: ten thousand “build an AI agent in 50 lines of Python” posts exist. “Here’s what happens when your Java agent hits production load week one” posts — almost zero. That gap is the whole bet. The DLQ fills up so you don’t have to explain to your CTO why the agent called the billing API 847 times. #Kafka #Java #LLM #SpringAI #SpringBoot #AIEngineering #BackendDevelopment

🚀 Java Deep Dive Series — Memory Management AI can write code. But understanding how memory works under the hood is what helps you write efficient and bug-free systems. Today, I revisited: 👉 Java Memory Management Here’s a quick breakdown 👇 🔹 Stack Memory → Method calls, local variables (thread-specific) 🔹 Heap Memory → Objects & shared data (managed by GC) 🔹 References → Strong, Weak, Soft (impact GC behavior) 🔹 Garbage Collection → Automatic memory cleanup 🔹 Generational Memory → Young → Old lifecycle ⚙️ Deep dive covered: Stack vs Heap with detailed examples, object lifecycle (Eden → Survivor → Old Gen), Minor vs Major GC, reference types behavior, Mark & Sweep algorithm, Metaspace (non-heap), different GC types (Serial, Parallel, G1, ZGC), and JVM tuning using -Xms & -Xmx. 💡 My Key Takeaway: Most performance issues in Java are not about logic — they come from how memory is used and managed. 📘 I’ve documented detailed notes (with examples & diagrams) here: 🔗 [https://lnkd.in/dsMypxEG] I’ll keep adding more topics as I go. If you're revising Java fundamentals or preparing for interviews, this might help 🤝 👉 Quick check: What kind of reference allows GC even when a reference still exists? #Java #JVM #MemoryManagement #GarbageCollection #BackendDevelopment #AI

Why Java is the Secret Weapon for Enterprise AI 🚀 Think AI belongs only to Python? Think again. While Python is great for experimentation, Java is becoming the first-class language for building AI at enterprise scale. Here is why Java is the future of the AI-powered enterprise: - Unmatched Runtime Efficiency: In the world of AI, every cycle counts. The JVM provides superior performance and efficiency compared to other runtimes. By saving budget on efficient execution, you can redirect those funds toward AI tokens and API calls - Enterprise-Grade Ecosystem: Java isn't starting from scratch. With frameworks like LangChain4j, Spring AI, and embabel, developers can seamlessly integrate LLMs and implement complex patterns like RAG and agentic flows using familiar tools - Context is King: AI needs data to be useful. Java has always excelled at integrating with third-party solutions, databases, and MCP servers, making it the perfect "integration layer" for providing AI with the necessary business context - Readability as a Superpower: As AI assistants (like GitHub Copilot and Claude Code) write more of our code, readability becomes more important than brevity. Java’s explicit nature makes it easier for developers to review and maintain AI-generated suggestions for critical apps With 62% of enterprises already using Java to power their AI applications, the "future" is already here. Java isn't just surviving the AI age; it’s providing the foundational execution layer for it What’s your take? Are you building your AI agents in Java, or are you sticking with Python for production? Let’s discuss in the comments! 👇 #Java #GenerativeAI #SoftwareEngineering #EnterpriseTech #JVM #SpringAI #TechTrends

🚀 Java Deep Dive Series — Interface AI can generate implementations. But defining the right contract between components is what makes systems scalable. Today, I revisited: 👉 Interfaces in Java Here’s a quick breakdown 👇 🔹 Interface Basics → Defines contract (what to do, not how) 🔹 Abstraction → Hide implementation, expose behavior 🔹 Polymorphism → Interface as a reference type 🔹 Multiple Inheritance → Achieved via interfaces (no diamond problem) 🔹 Default & Static Methods → Java 8 enhancements 🔹 Functional Interfaces → Single abstract method + lambda support ⚙️ Deep dive covered: Interface definition rules, fields (public static final), method rules, implementation in classes, nested interfaces, interface vs abstract class differences, default method conflicts & resolution, private methods (Java 9), functional interfaces, lambda expressions, and built-in functional interfaces (Consumer, Supplier, Function, Predicate). 💡 My Key Takeaway: Interfaces are not just for abstraction — they are the foundation of loosely coupled and extensible systems. 📘 I’ve documented detailed notes (with examples) here: 🔗 [https://lnkd.in/dEJb7gin] I’ll keep adding more topics as I go. If you're revising Java fundamentals or preparing for interviews, this might help 🤝 👉 Quick one: How does Java resolve the diamond problem with default methods? #Java #OOPS #Interfaces #FunctionalProgramming #BackendDevelopment #AI

I've been building RAG pipelines in Java for 12 months. Here's everything I wish I had on Day 1 — condensed into a free 10-page PDF guide. 📄 What's inside: ✅ RAG Architecture explained (Ingestion + Query phases) ✅ Spring Boot + LangChain4j full project setup ✅ Document chunking strategy (the #1 thing most guides skip) ✅ pgvector integration with HNSW index ✅ LLM augmentation with Claude — grounded, citation-aware answers ✅ Production patterns: Hybrid Search, Re-ranking, Metadata Filtering ✅ Complete REST API — curl examples included --- The hardest part of RAG isn't the LLM call. It's the retrieval layer. Most developers spend 90% of their time on prompt engineering and 10% on retrieval quality — and then wonder why their RAG chatbot hallucinates. The truth: garbage in, garbage out. Poor chunking → irrelevant chunks → the LLM makes stuff up. What actually moves the needle: → Recursive chunking (paragraph → sentence → word fallback) → 10-15% overlap between chunks to avoid cutting mid-thought → minScore threshold (≥ 0.70) to discard low-relevance matches → Metadata filtering for multi-tenant safety → Hybrid Search: vector + BM25 + RRF fusion Once you fix retrieval, the LLM answer quality jumps dramatically. --- Stack used in this guide: → Java 21 + Spring Boot 3.2 → LangChain4j 0.31 → Claude Sonnet (Anthropic) — 200K context window → pgvector on PostgreSQL — zero extra infra → OpenAI text-embedding-3-small --- 12 years of Java backend engineering + 12 months of GenAI production work. This guide is the bridge between the two. Grab the PDF above. Build something real. --- ♻️ Repost if this helps your team. 🔔 Follow @chandantechie for weekly Java + AI deep-dives. #Java #SpringBoot #RAG #GenAI #LLM #LangChain4j #AIEngineering #SoftwareEngineering #BackendDevelopment #VectorDatabase

Day 3 of Java with DSA Journey 🚀 📌 Topic: Guess Number Higher or Lower (LeetCode 374) 💬 Quote: "Efficiency is not about doing more; it's about eliminating what doesn't matter." ✨ What I Learned: 🔹 Binary Search Beyond Arrays: Binary Search isn’t limited to arrays — it works perfectly on a number range like [1...n]. 🔹 Working with APIs: Learned how to adapt logic based on API responses: -1 → Guess is too high 1 → Guess is too low 0 → Correct answer 🔹 Power of Efficiency: Even for a huge range (up to 2³¹ - 1), Binary Search finds the answer in ~31 steps 🤯 Compared to Linear Search → practically impossible! 🔹 Complexity: ⏱ Time: O(log n) 📦 Space: O(1) 🧠 Problem Solved: ✔️ Guess Number Higher or Lower 💡 Key Insight: This problem highlights the “Narrowing the Search Space” concept. Each step eliminates half the possibilities — that’s the magic of logarithmic algorithms ⚡ ⚡ Interview Insight (3-Way Decision Logic): Unlike boundary problems, here we deal with three outcomes: 1️⃣ 0 → Found the number (return immediately) 2️⃣ -1 → Move right = mid - 1 3️⃣ 1 → Move left = mid + 1 👉 Use while (left <= right) since the target is guaranteed to exist. 🔑 Takeaway: Consistency beats intensity. Showing up daily is what builds mastery. #DSA #LeetCode #Java #CodingJourney #BinarySearch #ProblemSolving #100DaysOfCode #JavaDeveloper #Algorithms

In this hands-on workshop, you’ll learn how to build production-ready AI agents in Java using LangChain4J - starting from a simple pipeline and evolving it into an agentic system with decision-making, guardrails, evaluation, and observability. By the end, you won’t just have an agent - you’ll understand how to control, measure, and debug it. Susanne Pieterse https://lnkd.in/d-T9Z26n

As agentic AI shifts from prototypes to enterprise production, Java emerges as a powerful alternative to Python-centric stacks. Here is My DZone article on 'Developing Agentic AI applications Using Java, LangChain4j, Quarkus, MCP, and OpenTelemetry for scalable enterprise apps' with code references! Like share and subscribe! 😃 This article looks into building robust agentic applications using LangChain4j for orchestration, Quarks for high-performance deployment, Model Context Protocol (MCP) for standardized tool and data access, and OpenTelemetry for comprehensive observability.. #BhaskarKollu #AgenticAI #MCP #RAG #ModelContextProtocol #GenerativeAI #DZone https://lnkd.in/gQAQzxMW