What if you tried to build a fully functioning data platform — ingestion, orchestration, analytics — entirely at home, on a single machine, with zero managed services? Not for work, not for a client, but as an after-hours experiment in self-learning, tinkering, and pushing your own limits. That’s the journey I embarked on: a nights-and-weekends project to create a platform that could tackle at least one real-world problem and deliver insights worthy of a decision.

In this talk, I’ll share the unfiltered story of that build — from picking the tools and wrestling with hardware constraints, to the experiments that succeeded, broke, or went sideways. We’ll explore the point where this made perfect sense for learning, the moment where reality caught up with ambition, and the trade-offs between “doing it right” and “just making it work.” The endgame: seeing the platform process real data and actually influence a decision, proving that curiosity projects can have tangible impact.