Hey there! 👋 I'm Stepanus Janu Adi Nugroho — a Backend Engineering undergraduate at Multi Data Palembang University, deeply passionate about building intelligent, automated, and scalable systems.

My current research focus is Automation Agent Systems — designing multi-agent architectures that can reason, orchestrate tasks, and collaborate across distributed pipelines. I'm especially fascinated by the intersection of AI engineering, backend infrastructure, and autonomous decision-making systems.

"The best systems are the ones that build themselves."

• 🔭 Currently researching: Autonomous agent frameworks, LLM-powered orchestration, and workflow automation
• 🌱 Learning: Agent memory architectures, tool-use patterns, ReAct & multi-agent coordination
• 💡 Passionate about: Simplifying complexity for mid-level engineers through clean system design
• 🎯 Goal: Building automation pipelines that reduce cognitive overhead in engineering workflows

My current deep-dive — designing systems where software agents think, plan, and act autonomously.

┌─────────────────────────────────────────────────┐
│           AUTOMATION AGENT ARCHITECTURE          │
│                                                  │
│  [Planner Agent] ──→ [Tool-Use Agent]            │
│        │                    │                    │
│        ↓                    ↓                    │
│  [Memory Store]    [Execution Runtime]           │
│        │                    │                    │
│        └─────────┬──────────┘                    │
│                  ↓                               │
│         [Validator / Reflector]                  │
│                  │                               │
│                  ↓                               │
│            Final Output ✅                        │
└─────────────────────────────────────────────────┘

Key areas of interest:

• 🧩 Multi-Agent Orchestration — Coordinating specialized agents (Recon, Planner, Executor, Validator)
• 🔗 LLM Tool-Use & Function Calling — Binding language models to real APIs, databases, and compute
• 🗃️ Agent Memory Systems — Short-term context, long-term vector stores, episodic memory
• ⚡ Workflow Automation — Replacing manual engineering pipelines with self-driving task graphs
• 🛡️ Reliability Engineering for Agents — Fault-tolerance, retry logic, output validation loops

Beyond backend systems, I'm captivated by technical animation as a communication tool — using motion to explain complex architecture to mid-level engineers at a glance.

Tools & approaches I explore:

• Mermaid.js — Live diagram-as-code for system flows
• Lottie / Rive — Lightweight animation for technical dashboards
• D3.js — Data-driven visual storytelling for system metrics
• CSS/SVG animation — Micro-interactions that make UIs feel alive

My philosophy: if a system is hard to explain in text, animate it.

 

graph LR
    A[Backend Core\nSpring · Laravel · Node] --> B[Agent Systems\nLLM Orchestration]
    B --> C[Multi-Agent\nCoordination]
    C --> D[Autonomous\nWorkflow Automation]
    D --> E[Production-Grade\nAgent Infrastructure]
    style A fill:#7C3AED,color:#fff
    style B fill:#2563EB,color:#fff
    style C fill:#059669,color:#fff
    style D fill:#D97706,color:#fff
    style E fill:#DC2626,color:#fff

"Complexity is the enemy of reliability. Automate the repetitive, architect the essential."

I build for middle engineers — systems and tools that are powerful enough to handle real-world scale, but clear enough that you don't need a PhD to maintain them. Good architecture tells its own story.