~/julio $ whoami
I'm Julio Chana, AI systems engineer. Everything here is a distillation of the most useful skills I refine in daily work: courses you can take, skills that teach coding agents new tricks, and real shipped projects.
Live workshops and courses I teach and sell. Cohorts are small on purpose. When they're full, they're full.
An in-company workshop for executives and teams: what AI actually does, live demos with your own use cases, and a prompt playbook your people keep using the next day.
For professionals drowning in information: capture, organize and retrieve everything you know with AI. A system you build once and use forever.
Turn repetitive business work into bots and flows: invoicing, follow-ups, reporting. Built for small teams that want results, not infrastructure.
Booking for your company or team? . They skip the waitlist.
Drop-in skills for coding agents like Claude Code: a distillation of the most useful workflows I refine in daily work. Install one and your agent learns a new trick.
Audits a site's metas, Open Graph, robots.txt and sitemap, then fixes what's broken instead of just listing it.
$ /seo-audit
Serves your site locally, drives a headless browser and screenshots every section, so the agent can see what it built before you do.
$ /ui-verify
Battle-tested prompt patterns for business automation, packaged so your agent applies them instead of you copy-pasting them.
$ /prompt-pack
Recipes for the automations small businesses ask for first: collections bots, invoice chasing, report generation.
$ /auto-flows
First public drops land on GitHub. Want early access? .
Projects and experiments. Some open, some shipped for clients, all real.
An interactive web app that serves inspiring quotes on demand. Small, fast and built for fun: the kind of weekend project that keeps the craft sharp.
A commissioned voice-synthesis demonstration: a cloned voice called a TV host on air to show, on Swedish national television, what AI fraud sounds like. Built for awareness, broadcast on TV4.
Watch the segment →Founding engineer on a real-time computer-vision SDK shipped to thousands of enterprise integrations. Drove per-frame inference from 300 ms down to 6 ms (a 98% reduction) and showcased it at CES, AWE and MWC.
manomotion.com →For Aclara's Tierras Raras de Penco project: machine-learning models that predict anodic effects during the metallization stage of rare-earth production from gas-emission telemetry, so the process team intervenes before problems happen, not after.
aclara-re.com →
An LLM + RAG pipeline that turns informal email requests into validated commercial quotes against a structured product database. End to end, in production for a real business.
More on github.com/jachana.
Everything else that doesn't fit a box.
~/julio $ ping
Course requests, project ideas, consulting, or just to say the voice-clone segment was scary. My inbox is open.
info@juliocode.com
