Stages of AI-Assisted Development: How My Workflow Evolved
A practitioner's account of how AI integration in software development evolved incrementally, from basic coding assistance to structured spec-first workflows and agent-based systems.
Principal Product Manager
Bridging innovation and business value. I build real tools, understand infrastructure, and ship code. 13+ years turning complex problems into products that work.
Thoughts on product, engineering, and AI
A practitioner's account of how AI integration in software development evolved incrementally, from basic coding assistance to structured spec-first workflows and agent-based systems.
Building is how I learn. Writing code forces me to deal with real constraints, trade-offs, and failures. I construct small tools and experimental projects publicly to understand systems deeply.
Exploring how AI adoption impacts skill development in younger generations. Potential cognitive losses like declining critical thinking alongside emerging capabilities like prompt engineering.
Real tools I've designed and built
A production-grade MCP server for Actual Budget with 62 tools, enabling AI assistants to manage budgets, transactions, accounts, and financial reports through natural language.
62 tools | Production-ready | Docker
Remote AI CLI control via Telegram and Slack with multi-agent orchestration. Control AI coding agents from your phone while they work on your infrastructure.
Multi-agent | Remote control | Real-time
Distributed VPN monitoring system with DNS leak detection, multi-node health checks, and real-time alerting. Ensures your VPN infrastructure stays secure.
DNS leak detection | Distributed monitoring