Nikola Balić
I build AI systems and write about what I learn.
What I Build
I ship AI experiments and production systems. Recent work:
- Looper – Autonomous coding system with review gates and deterministic workflows
- AgentProbe – Open-source tool for testing AI agent CLI interactions
- Agentic AI Handbook – 113 production-ready patterns for multi-agent systems
- This site – Built and maintained with AI coding agents
I've shipped 50+ AI projects in the past year, documenting what actually works in production.
Current Focus
- Multi-agent orchestration – How autonomous agents coordinate at scale
- Agent-friendly infrastructure – Type safety, CLIs, machine-readable docs
- Performance patterns – 5x improvements through strategic coordination
- AI-native development – What changes when AI is your primary collaborator
Writing & Research
I write detailed technical analysis:
- "The Agentic AI Handbook" – 113 battle-tested patterns
- Infrastructure deep-dives – Serverless costs, caching strategies
- AI development patterns – What Sourcegraph learned, bottleneck analysis
- Technology adoption – 20-year cycles and AI's acceleration
Everything at /log – no fluff, just detailed technical writeups.
Work
I've worked with:
- Daytona – AI runtimes for agentic workflows, open source tooling
- Codeanywhere – Cloud development environments
- Steel – Developer infrastructure
- Numarics – Early-stage fintech
Growth roles, finding PMF through PLG and GTM strategies for developer tools and AI infrastructure.
Background
- Lecturer in Data Science – NoSQL, Complex Network Analysis, SWE 101, Scientific Programming
- Doctoral research – Digital Transformation (published academic work)
- 14 years in tech – Started as engineer, stayed technical
For formal credentials: /bio
Right Now
Experimenting with autonomous agents that run for hours without human intervention, content delivery systems that serve both humans and AI agents, and performance optimization for AI-native workflows.
I write about what works, what doesn't, and why.