Hola ✨ I’m Mus, a Senior Software Engineer at Platzi, where I design and operate LLM-powered educational systems that evaluate every student on the platform.

I work at the intersection of product, instructional design, and prompt & context engineering. My path started in frontend. I spent five years building Platzi’s product surface before moving into AI engineering full-time.

Get in touch

Case study

Challenges

An LLM pipeline that auto-generates gamified, interactive learning exercises and embeds them at the right moments in videos and formal exams.

Owner & developer · Platzi · 2024–present · Live

Problem

Interactive checkpoints had to be authored by hand, per video. That does not scale across thousands of hours of content: authoring effort grows with every class, quality drifts between authors, and students mostly watched passive video with few comprehension checks.

What I built

  • Owned the full prompt layer: base prompt, type-specific prompts, cognitive-demand hierarchy, and autonomous mode selection.
  • Split generation (creator) from timestamp placement (positioner) inside VTT transcripts.
  • Ran Langfuse prompt versioning and LLM-as-judge evals; built React/CSV audit tooling to flag deficient challenges.

Key fix

Early outputs were too easy. Added a cognitive-demand hierarchy and constraints across base and type-specific prompts, plus an audit pass to catch trivial exercises.

Results

98%

Completion rate in video classes with Challenges

100%

Platzi courses with Challenges in every class

Live

Interactive questions in formal exams

ClaudeLangfusePythonElevenLabsReact

Want the technical breakdown? Get in touch

Last work

Frontend engineer turned prompt engineer, still shipping product. Here’s where I’ve built systems that reach real students, not just impressive demos.

Full time

Senior Software Engineer

Design and operate LLM-powered educational systems: diagnostic placement assessments, gamified Challenges inside videos and exams, skill extraction from course transcripts, and challenge positioning in VTT transcripts with LLM-as-judge evaluation. Established prompt evaluation workflows in Langfuse with structured JSON schemas and quality metrics.

Python Langfuse Claude API Next.js

Full time

Front End Developer

Led migration from micro-frontends to a Next.js 13 monorepo. Shipped Google SSO Login (67% of platform logins), a checkout redesign that lifted mobile conversion +1.7%, and Redeemable Codes & Coupons for HP/Lenovo partnerships (50K+ codes). Taught a Platzi course on responsive mobile web development.

React Next.js TypeScript Monorepo

Full time

Software Engineer

Designed and built software services, mobile apps, and UIs from concept to production for a bulk-purchasing platform. Stack: Node.js, React, React Native, AWS, Serverless Framework.

React React Native AWS Node.js

I put myself out there

Speaker at PlatziConf on AI and applied LLM work. Creator of @musarte.dev: technical content on AI, prompt engineering, and web development for Spanish-speaking developers.

I'm a Senior Software Engineer at Platzi, Latin America's largest tech education platform. I design and operate the LLM-powered systems that generate and evaluate the educational content students interact with every day: placement tests, interactive challenges, self-assessments, exam questions, and the skill taxonomy underneath them all.

My path to AI engineering came through frontend. I spent five years building Platzi’s product surface (micro-frontend architecture, checkout, growth, B2B) before moving into prompt and context engineering full-time. That product foundation is what lets me build AI systems that actually serve real students, not just impressive demos. Outside of work, I teach AI, prompt engineering, and web development in Spanish as @musarte.dev on Instagram.

Skills

  • AI / LLM engineering: prompt & context engineering, LLM evaluation, Langfuse, structured JSON output, RAG, LLM-as-judge patterns, Claude API
  • Languages & frameworks: Python, JavaScript/TypeScript, React, Next.js, Node.js, React Native
  • Infrastructure & data: AWS Lambda, Step Functions, Serverless, Cloudflare Workers, SQL, MongoDB, Firebase, REST APIs
import Developer from "./Developer";

class Mus extends Developer {
  constructor() {
    super();
    this.role = "Senior Software Engineer";
    this.company = "Platzi";
    this.stack = ["Python", "TypeScript", "Langfuse"];
    this.teachesAs = "@musarte.dev";
  }

  buildChallenge(course) {
    return this.llm.generate({ course, type: "mcq" });
  }

  shareWhatILearn(topic) {
    return this.instagram.post(topic);
  }
}

export default new Mus();

Bonus track

👾

People started to call me Mus more than 15 years ago!

🎤

Speaker at PlatziConf on AI and applied LLM work, and to corporate leaders on LLMs, RAG, and AI literacy.

📱

Creator of @musarte.dev on Instagram, one of the few Spanish-language technical brands focused on applied LLM work.