Hi, I'm

Pol Casacuberta

I build AI & data systems that work in the real world.

Computer Science Engineer in Barcelona and AI & Data Consultant at Deloitte. I turn ideas into production-ready systems across cloud data platforms, computer vision, and generative AI.

View my work
Professional Work

Featured Projects

Key milestones and AI initiatives delivered in the real world.

Client:Horse

Physical AI Anomalies Detection

AI & Data Consultant @ Deloitte

The world's first Physical AI project delivered at Deloitte. Built on NVIDIA Omniverse, Isaac Sim, and the Metropolis stack (DeepStream, TAO, Triton) to detect manufacturing anomalies on live motor production lines in real time, plus a warehouse digital twin with camera-based package tracking.

  • Physical AI
  • NVIDIA Omniverse
  • Isaac Sim
  • Metropolis
  • Digital Twin
  • Real-time
Client:Gobierno de Aragón

E2E AI Talent Matching Platform

AI & Data Consultant @ Deloitte

End-to-end software that parses PDF resumes, standardizes them into structured JSON, and indexes them in a RAG database. Allows semantic querying to find the best candidates for a job description and vice versa.

  • LLMs
  • RAG
  • Vector Search
  • PDF Parsing
Interactive

Try my projects

Live, interactive demos you can play with right in your browser.

Live Demo

Bachelor's Thesis — Polyp Detection

Bachelor's thesis on detecting polyps in colonoscopy images with deep learning, using CycleGAN and SPADE to generate synthetic training data when real labels were scarce.

  • PyTorch
  • CycleGAN
  • Object Detection
  • UPC
Live Demo

Draculin

Health-tracking app with an AI chatbot and computer vision assistance, built in Flutter and Django as a BitsXLaMarató 2023 finalist.

  • Flutter
  • Django
  • AI
Live Demo

BitsXLaMarató — Aorta

BitsXLaMarató 2022 winner: a tool to segment the aorta in ultrasound videos with Mask R-CNN and reconstruct it in 3D to help detect aneurysms.

  • PyTorch
  • OpenCV
  • 3D
  • CV
Live Demo

License Plate Detection

Computer vision project to locate and read license plates from car photos, originally built as a classical MATLAB pipeline.

  • Computer Vision
  • TypeScript
Live Demo

Programming Game Viewer

Data structures course project: a grid-based game where C++ bots compete as warriors and builders in a class-wide tournament.

  • Mithril.js
  • Canvas
  • C++
  • FastAPI
Live Demo

JSBach Interpreter

Compilers course project: JSBach, a small programming language where you describe musical pieces as code, built with an ANTLR4 grammar.

  • ANTLR4
  • Web Audio
  • Language
Live Demo

Tenda Online

Web applications course project: a full e-commerce store built in PHP and MySQL, covering product catalog, shopping cart, and checkout flow.

  • PHP
  • MySQL
  • E-Commerce
Live Demo

WPGMA Clustering

Programming course project: implementing WPGMA clustering in C++ to build phylogenetic trees from species gene sequences.

  • TypeScript
  • Algorithms
  • D3
Live Demo

MPIDS Solver

Algorithms project on the Minimum Positive Influence Dominating Set problem, solved with greedy heuristics and simulated-annealing-style local search.

  • Elm
  • D3.js
  • FastAPI
  • Graph Theory
Live Demo

Graph Phase Transitions

Algorithms research project studying connectivity and complexity phase transitions on binomial, geometric, and grid random graphs.

  • Lit
  • Canvas
  • FastAPI
  • NetworkX
Live Demo

Travel agency (PDDL)

Automated planning course project: a PDDL domain for booking trips with flights, hotels, numeric constraints, and metric minimization.

  • PDDL
  • Planning
  • UPC
Live Demo

Desastres IA

AI course project framing disaster relief as local search, with helicopters rescuing groups from rescue centers via hill climbing and simulated annealing.

  • Solid.js
  • FastAPI
  • Local search
  • Canvas
  • UPC
Live Demo

Machine Learning — Hypothyroid

Machine learning course practica: training and comparing nine sklearn model families to predict hypothyroidism from clinical data.

  • Python
  • Jupyter
  • sklearn
  • UPC
Live Demo

Recommendation System

Software engineering course project: a content-based recommendation system built in Java with Spring Boot and Thymeleaf, including a full web interface.

  • Java
  • Spring Boot
  • Recommendation
Live Demo

CAIM IR Explorer

Information retrieval course labs: computing PageRank over the global airport network and verifying Zipf's Law on text corpora.

  • Vanilla TS
  • D3.js
  • FastAPI
  • PageRank
Live Demo

Trip Planner (Expert System)

Knowledge-based systems course project: an expert system trip planner originally written with CLIPS production rules, reimplemented in Python.

  • HTMX
  • Alpine.js
  • Litestar
  • Expert System
Live Demo

Parallel Computing Labs

Parallel computing course labs in C with OpenMP: Mandelbrot, heat equation, 4-way merge sort, π estimation, and speedup benchmarks.

  • C
  • OpenMP
  • WebAssembly
  • Preact
  • Canvas
Live Demo

Robotics Dashboard

Robotics course labs covering odometry, wall-following, 3D arm forward kinematics (Puma560 / 3-Link), trajectory planning, sensors, and EKF state estimation.

  • Ember.js
  • Babylon.js
  • Robotics
  • Canvas
  • EKF
Live Demo

Algorithm Visualizer

Algorithms and data structures coursework covering 18 classical algorithms across graphs, sorting, search, and combinatorial problems.

  • Qwik
  • Canvas
  • Algorithms
  • Graphs
  • Sorting
Live Demo

Computer Graphics Shader Playground

Computer graphics course labs originally written against desktop OpenGL, covering vertex transforms, Phong lighting, procedural textures, fragment patterns, and geometry shader emulation.

  • WebGL2
  • GLSL
  • Shaders
  • UPC
About

A bit about me

Hi! I'm Pol — a Computer Science Engineer from Barcelona. I graduated from UPC's Facultat d'Informàtica de Barcelona (FIB) and now work as an AI & Data Consultant at Deloitte, where most of my days are spent at the intersection of AI, data platforms, and the messy reality of making them useful.

What I actually enjoy is building stuff — exploring new tools, figuring out how they fit together, and helping my team ship things we're proud of. I care more about solving the problem than about picking the trendiest stack, though I'm happy when both line up.

Outside of work you'll usually find me climbing, deep in a videogame, or pretending my Vim keybindings make me faster than I really am. Here's some of the tech I've been using lately:

  • Python
  • SQL & Power BI
  • AWS & Azure
  • Machine Learning
  • Generative AI
  • Computer Vision

Download CV

CV version
Balanced version with profile, experience, education, extracurricular activities, projects, and skills.
Download
Pol Casacuberta
Experience

Where I've worked

A timeline of my professional journey so far.

  • Sep 2024 — Present ·

    AI & ML Engineer @ Deloitte

    • Delivered Deloitte's first Physical AI project: real-time manufacturing anomaly detection on live motor production lines and a warehouse digital twin with camera-based package tracking, built on NVIDIA Omniverse, Isaac Sim, and the Metropolis stack (DeepStream, TAO, Triton)
    • Designed and deployed AI and data infrastructure across AWS, Azure, and GCP — edge inference, streaming pipelines, and cloud-hosted training and model serving
    • Embedded in the Generalitat de Catalunya's Data Governance Office (OGD) to help define the institution-wide data strategy — established data catalogues, mapped asset lineage with Informatica, and delivered governance advisory to cross-departmental teams across the public administration
  • Sep 2023 — Sep 2024 ·

    AI Engineer @ MGS, Seguros y Reaseguros S.A.

    • Built pricing prediction models as part of a 4-person AI and data team supporting insurance product analytics
    • Automated SQL Server workflows with PowerShell to streamline recurring data processing and reporting tasks
    • Developed geospatial analysis projects in R / RStudio, using map data to support location-based business insights
Skills

Tech I work with

Technologies and tools I use to bring ideas to life.

Programming Languages

Python]TypeScript]JavaScript]C]C++]Java]SQL]

AI & Data Science

Machine Learning]Computer Vision]PyTorch]scikit-learn]OpenCV]LangChain]Object Detection]Generative Models]MLOps]Data Mining]Statistics]

Tools & Platforms

AWS]Azure]GCP]Docker]GitHub Actions]PostgreSQL]FastAPI]Git]Jupyter]

CS Fundamentals

Algorithms]Software Architecture]Parallel Processing]CI/CD]Compiler Design]Cybersecurity]
Credentials

Licenses & certifications

Professional certifications I have earned along the way.

Education

Where I studied

Contact

Get in touch

I'm currently open to new opportunities and collaborations. Whether you have a question, a project idea, or just want to say hi — my inbox is always open.