About me
I'm Juan Torchia — an Argentinian Full Stack developer specialized in software architecture and distributed systems.
I work with Next.js, TypeScript, Java, and Spring Boot. I'm obsessed with performance, DX, and shipping AI into real products.
I write here to share what I learn while building. In Spanish and English, so it reaches as many devs as possible.
What I write about
Latest articles
See all 240 →I ran OpenAI's security audit 10 times: to review your own code, plain ChatGPT is enough
Everyone vibecodes and then wants to review their own code, and OpenAI says that's what Daybreak Blue is for. I ran the same security audit with plain ChatGPT and with Blue, on the same real codebase, ten times, to see whether the special tier finds anything the regular model doesn't.
ExperimentsgRPC shows up when REST starts breaking under its own weight
gRPC solves inter-service communication with code generation and HTTP/2. It's fast, it's typed, and the learning curve is going to hurt a bit. Spoiler: it's worth it.
ExperimentsApache Spark won't melt down when your data doesn't fit in RAM anymore
Apache Spark has been the de facto standard for processing data at scale for over a decade. We go over why it's still alive, how it works under the hood, and when you should NOT use it.
ExperimentsI Gave My VS Code Extension to a Cybersecurity Model and It Found an Infinite Hang
Daybreak Blue audited CertView, my own published plugin. It found two ways to freeze VS Code with a file smaller than a photo. One of them never ends.
TutorialsDocker HEALTHCHECK: what the command actually runs and what it doesn't tell you
HEALTHCHECK in Docker runs a specific command and looks at its exit code. It doesn't know if the database responds, or if an external dependency is down. That difference matters when a container stays "healthy" while the app is broken inside.
TutorialsMem0 Doesn't Fix an Unbounded Agent, It Complements It
I designed a small experiment to see if Mem0 reduces context bloat in long Cline sessions. Short answer: it helps with repetition, it doesn't replace limiting the agent's scope.