Writing on Medium & LinkedIn.
I write about what I’m learning, building, and thinking about across AI engineering, software systems, and product development. I share lessons, experiments, architectures, and perspectives from building technology in the real world.
Engineering & system design
Event-driven architecture, analytics pipelines, scaling read-heavy systems, and modernizing legacy code without breaking production.
Leadership & execution
Aligning teams around shared outcomes, execution rhythm, and the difference between activity and impact.
Product thinking
Why onboarding beats feature count, diagnosing before building, and turning business goals into technical roadmaps.
The shift to AI
Applied generative AI — RAG, agents, MCP — and what it takes to build reliable, production-grade intelligent systems.
Stop Building Multi-Agent Systems Just Because They’re Trendy
Multi-agent systems are powerful, but not every AI product needs them — a practical framework for deciding when to scale one workflow vertically, and when to scale agents horizontally.
Read on Medium →
The Trust Problem I Realized While Planning My Recent Project
Connecting external MCP servers to AI systems expands your trust boundary — why developers need to evaluate third-party integrations carefully rather than adopting them for convenience.
Read on Medium →
Architecture Before Tools
Architecture is the fundamental design based on business requirements and cost constraints — identify the problem first, then pick the tools, not the other way around.
Read on Medium →
Horizontal AI Platforms vs Vertical AI Products: Two Mindsets Every AI Engineer Should Understand
Building a reusable horizontal platform versus a vertical product that solves one business problem exceptionally well — a decision that shapes architecture, priorities, and how value gets delivered.
Read on Medium →
From Prompt Engineering to Graph Engineering
How prompt, context, harness, loop, and graph engineering each solve a different layer of AI agent systems — walked through with a refund-assistant example.
Read on Medium →
Don't Graph Everything
A practical framework for deciding when a system actually needs Prompt, Context, Harness, Loop, or Graph Engineering — architecture should match the problem, not the trend.
Read on Medium →
My Experiment with Turborepo, AI, and Modern Monorepo Thinking
How Turborepo streamlines JavaScript/TypeScript monorepo management, and how AI-assisted development reshaped my thinking about code organization.
Read on Medium →
Engineering Notes from Building a Revenue Intelligence System
Key decisions and tradeoffs from building a revenue intelligence system — good engineering isn't about applying best practices blindly.
Read on Medium →
What I Learned Building a Revenue Intelligence System
Don't build systems that show data — build systems that help people decide. Lessons on data modeling, clear metrics, and constraint-based design.
Read on Medium →
Building an Intelligence Layer: From Data to Decisions
Clean data gives answers, intelligence gives direction — bridging the gap between data analysis and actionable decision-making.
Read on Medium →
Why Data Modeling Matters More Than Dashboards
Why clean data isn't enough — and how defining meaning changes everything in your system.
Read on Medium →
Why CSV Ingestion Pipelines Are Harder Than They Look
The unexpected complexities that emerge when building data ingestion systems from CSV files — challenges that go well beyond simple parsing.
Read on Medium →
How I Built a Revenue Intelligence System from Monthly CSV Data
Building an automated decision-support system that transforms monthly billing data into actionable revenue insights, moving beyond traditional dashboards.
Read on Medium →If You Don't Read in the AI Era, You'll Slowly Lose Your Mind
Reading — especially books — is essential for imagination, critical thinking, and cognitive ability in an era where AI provides quick answers but can't build human mental capacity.
Read on Medium →
We Built an SMS Feature… and Discovered a Hidden System Design Flaw
How a system design flaw let users consume costs faster than subscription revenue was collected — and what it taught us about designing for the failure case, not just the happy path.
Read on Medium →The Evolution of Programming Languages
How technology shifted from humans learning machine code to machines learning human language — programming's path from binary to natural language interfaces.
Read on Medium →I Tried a Structured AI Development Workflow in Cursor
Experimenting with a systematic approach to AI-assisted development in Cursor that kept project context across features — until token usage from repeated caching got out of hand.
Read on Medium →From Code Craftsman to AI Orchestrator: My Journey as a Software Engineer in the AI Era
How AI tools transformed my role from hands-on coding to system orchestration — accelerating development while demanding deeper architectural thinking.
Read on Medium →
The $37B Meeting Problem—And How We Solved It
Ineffective meetings cost companies billions a year — a systems-based approach to fixing meeting culture and getting time back.
Read on LinkedIn →
From Chaos to Clarity: How We Ran Our Most Productive Workshop
How a structured, all-day design-thinking workshop turned a struggling team's meeting culture around.
Read on LinkedIn →Being a Software Engineer in the Era of AI: Replacement or Reinvention?
Is AI replacing software engineers or reshaping the role? What AI does best, where humans still win, and where software engineering is headed.
Read on Medium →Born In 2025? Welcome To The Beta Age.
An exploration of Generation Beta (born 2025–2039) and how they'll grow up in an AI-integrated world shaped by technology and climate change.
Read on Medium →