AI Engineer·Full-Stack Developer
Building at the intersection of intelligence and real-world systems.
Available for work
AGS, MX — 2026
About
Story
I'm an AI Engineer pursuing a postgraduate specialization in Engineering & Project Management at Universidad Panamericana. Currently serving on the Student Council, I build full-stack platforms, from automation pipelines to production web apps, that real people actually use.
Aug 2026Present
Specialization in Engineering and Project Management
Universidad Panamericana
Focused on end-to-end technical project lifecycles, resource management, and Agile/Scrum delivery.
Jan 2026Present
Universidad Panamericana Student Council
Active Member
Active member of the student council at Universidad Panamericana, collaborating on student initiatives, representation, and campus improvement efforts.
Jan 2025Aug 2026
Foráneos UP
President
Guided and supervised team operations, promoted collaboration, and fostered an efficient work environment aligned with objectives.
Aug 2024Dec 2024
Burkle Exterminadores
Web Developer
Developed and deployed the company website to attract clients, designed a blog and Q&A forum to increase interaction, and optimized structure for better navigation.
Aug 2023Dec 2024
Integra CEF (Foreign Students Commission)
Vice President
Supervised strategic functions of team members, ensured completion of responsibilities, and developed the organization's website.
Aug 2022Aug 2026
Artificial Intelligence Engineer
Universidad Panamericana
Study of artificial intelligence techniques and algorithms for building intelligent systems.
Stack
Core
Python C++ JavaScript TypeScript
AI & Data
PyTorch n8n Apify SQL
Web & Platforms
React Firebase Supabase Vite Astro Next.js TailwindCSS Stripe
Selected Work
Projects I've shipped
Scroll through each one
Featured
01 / 05
01 — web
UPMIND 2026
Event Ticketing Platform
A ticketing and access platform for UPMIND 2026. Attendees get a QR ticket generated automatically after payment, and staff scan them at the door with live check-in tracking instead of a printed list. Receipt verification runs automatically too, so nobody's manually cross-checking payments against a spreadsheet on event day.
Notaria 71's public website: service pages, a contact flow, and an information structure built around trust. It's a notary's office, so credibility matters more than flash. Since launch it's been the main channel bringing in new client inquiries.
A CRM and billing system for a jewelry business, with Stripe handling payments and Supabase edge functions running the automation behind invoice generation and client follow-ups. Before this, invoicing was manual. Now it's one less thing the owner has to think about.
React Supabase Stripe Automation
04 — web
Collector's Vault Store
E-commerce Website
An online store for Collector's Vault, built on Next.js with Shopify Headless. Catalog browsing and checkout are tuned so the buying flow doesn't lose people partway through.
Rebuilt Foráneos UP's community site from a legacy CMS to Astro, mainly to fix how slow it had gotten. Now it's where out-of-state students find housing listings, relocation guides, and campus events in one place instead of scattered across group chats.
Algorithm implementations, computer vision experiments, embedded systems,
and everything else that doesn't fit neatly into a case study.
Web App
Flux
Flux is a personal finance app for tracking accounts, cards, and cash flow in one dashboard. I built it because I wanted something that also handled splitting expenses with roommates properly, with its own settlement logic instead of a spreadsheet everyone forgets to update.
An event ticketing and fundraising platform with tiered pricing and payment processing built in. Each purchase triggers a QR ticket by email automatically, and staff get a check-in dashboard to validate attendance live at the door.
Vite Next.js Supabase
web
Foráneos UP Dashboard
An operations dashboard for Foráneos UP's internal team, built on React and Firebase. It handles bulk data imports, tracks changes over time, and surfaces live analytics so the team isn't coordinating everything over spreadsheets and group chats.
React Firebase Material UI
web
Housing Dashboard
An admin dashboard for moderating and publishing verified student housing listings, with role-based permissions so different staff only see what they need, plus automated approval flows and status tracking to keep listings accurate as properties come on and off the market.
Vite Supabase
web
ToSignal
ToSignal automates LinkedIn lead research. n8n and Apify handle the extraction and processing, Supabase stores it, and the dashboard updates close to real-time with PDF reports generated automatically. What used to be hours of manual lead research per workflow now mostly runs itself.
React Supabase n8n
web
Influencer Data & Automation Web Platform
A large-scale influencer analytics platform, still in progress. n8n and Apify run the data pipelines, and SQL workflows in Supabase process high-volume datasets that used to be aggregated by hand.
n8n Supabase SQL
web
Garbon Albercas
Garbon's business website, built mobile-first since most of their traffic comes in on phones. Service info is laid out to get visitors to the contact form quickly.
DeepReading turns raw reading material into study material. An LLM generates summaries and flashcards automatically, so turning a long article or document into something you can actually review takes minutes instead of doing it by hand. Built with React, Vite, and Supabase.
React Vite Supabase
web
Burkle Fumigaciones
Burkle Fumigaciones' business website: clear service pages and a contact flow built to turn visitors into inquiries, which is really all a site like this needs to do well.
A CNN-based pipeline for detecting Mexican banknotes, built in PyTorch with OpenCV for preprocessing and augmentation. The goal was making classification hold up under real-world visual variance: different lighting, angles, wear on the bill, not just clean training images.
An end-to-end medical imaging pipeline using ResNet18 and DenseNet121, with CLAHE preprocessing and Grad-CAM so the model's predictions are actually interpretable, not just a black-box accuracy number. Hit 88.75% accuracy on the eval set.
A handwritten-digit recognition pipeline (preprocessing, training, evaluation loops) built as a baseline I could reuse for later computer-vision experiments rather than starting from scratch each time.
An image classifier for identifying animal species, developed through iterative notebook experimentation. Mostly a chance to get faster at picking the right model and tuning it with actual data instead of guesswork.
A working oscilloscope on an STM32F401RE: ADC sampling, buffering, and real-time waveform rendering, done in embedded C. A cheap way to inspect signals when you don't have a real scope on hand.
A 3x3x3 Rubik's Cube solver comparing BFS, A*, and IDA* against each other, with visualization tooling to actually watch how each algorithm searches the state space and see why one finds a shorter solution than another.
A Quixo-playing bot using minimax with alpha-beta pruning and hand-tuned heuristics. The interesting part was getting it to still play well under a tight per-move time limit, not just given unlimited thinking time.
A B-Tree implementation in Python: insertion, balancing, and indexed search from scratch, mostly to actually understand why databases use this structure instead of just taking it for granted.
A small numerical computing toolkit in Python covering approximation, interpolation, and equation-solving, built to have a reusable, tested set of methods instead of rewriting the same math from scratch for every assignment.
A Huffman-coding file compressor: encode and decode, with I/O handled efficiently enough that it's actually usable on real files, not just toy examples. Cut file size by up to 20% in testing.
A regex construction and testing toolkit with real-time pattern validation, built because debugging a bad regex by trial-and-error in production code is miserable. This lets you iterate on the pattern before it ever touches real text.
A Mastermind game engine in C++, with deterministic feedback logic and rule validation on every guess. Built as an exercise in getting the game logic exactly right rather than approximately right.
