CertiShot
Cryptographically certified photo evidence.
- React Native
- FastAPI
- PostgreSQL
- Docker
- Python
2026
Software Engineer
Computer Science student at UAB. Passionate about building full-stack systems and developing useful AI tools, always looking for ways to integrate smart features into solid backends.
About Me
Computer Science student passionate about technological innovation. I define myself as a curious, self-taught, and adaptable person with a great ability to collaborate in a team. I enjoy designing and developing personal projects that require planning and creativity.
My interest in 3D printing, nature, and sports provides me with a problem-solving mindset, allowing me to bring different approaches to the projects I participate in.
Languages
Catalan
Native
Spanish
Native
English
C1
German
Elementary
Recent Projects
Price Tracker
Modular e-commerce price monitoring.
- Python
- FastAPI
- SQLAlchemy
- Streamlit
- Docker
2025
AI Code Reviewer
Automated multi-language code review with Llama 3.
- Python
- Llama 3
- Groq
2025
Acorde
Auto-generated guitar tablature.
- Python
- Swift
- Dart
- Flutter
2026
Helicopter Game
Classic arcade helicopter clone.
- Python
2026
Tetris
Tetris built from scratch in C/C++.
- C
- C++
2026
Tools I Work with
Backend
- Python
- FastAPI
- Java
- C / C++
- PHP
- REST APIs
Frontend
- JavaScript
- Astro
- Tailwind CSS
- React Native
- HTML / CSS
Infra & Data
- Docker
- PostgreSQL
- MongoDB
- Git / GitHub
- Cloudflare
- Linux
Ask my AI anything
Meet my custom AI assistant, built to help you explore my portfolio. You can ask it anything about my tech stack, development experience, or university background.
Ask anything about my background, projects or skills.
How this chat works ▾
-
01
Static frontend, interactive island
The site is a static Astro build. Only this chat is a hydrated Preact island (loaded on scroll), so the page stays fast and SEO-friendly.
-
02
Server-side proxy holds the key
A FastAPI service holds the OpenAI key and talks to the model. The browser never sees it — the frontend only calls a same-origin /api/chat endpoint.
-
03
Curated context, not a vector DB
My bio, projects and skills fit in the prompt, so the assistant uses curated context injection instead of RAG — simpler, cheaper, and the right tool at this scale.
-
04
Rate limiting protects the budget
Per-IP limits, a global daily cap and a capped response length keep costs bounded and prevent abuse from draining the API budget.