Pol Marsol logo
SOFTWAREDEVELOPER
AIMACHINELEARNING
Igualada, Barcelona

Aboutme

Open to opportunities

I'm a 4th-year Computer Engineering student at Universitat de Lleida, passionate about applying AI and Machine Learning to solve real problems.

During my internship at ENGISOFT, I built ML models for clinical diagnostics and industrial anomaly detection, deployed within a web dashboard with Flask and REST APIs for real-time inference.

I'm currently developing NECODYS — my bachelor's thesis — a clinical mobile app for early dyslexia detection in children aged 3–6, in collaboration with the IRB research group.

I'm looking for opportunities where AI makes a real difference.

Pol Marsol

Experience

Oct 2025 – Present

Full-time

Software Developer – AI & ML

ENGISOFT Cloud Services

  • Development of ML and AI models applied to medical diagnostics and industrial anomaly detection.
  • Clinical web dashboard with Flask backend and REST APIs for interactive real-time inference.

Jul 2024 – Aug 2024

Internship

Full Stack Developer

BonÀrea

  • Full development (HTML frontend + Java backend) of a lottery management module for the corporate intranet across all BonÀrea stores, including business logic and validations.

Projects

Doctor examining a chest X-ray in a clinical settingReal photo
ENGISOFT · 2025
01FEATURED

Pleural Disease ML Diagnosis

Multiclass classification system (Heart Failure, Cancer, Parapneumonic, Tuberculosis) on clinical tabular data. XGBoost, LightGBM and CatBoost optimised with Optuna. Clinical explainability with SHAP and One-vs-Rest architecture.

PythonXGBoostLightGBMCatBoostSHAPOptunaScikit-learn
Private (internship)
Workers inspecting textiles in a factory quality-control environmentReal photo
ENGISOFT · 2026
02

Tissue Anomaly Detection via Autoencoders

Convolutional U-Net autoencoder for unsupervised defect detection in industrial fabrics using reconstruction error heat maps (MSE). Patch pipeline with data augmentation on GPU, sliding window inference.

PythonKerasU-NetAutoencodersOpenCVGPU
Private (internship)
Child using a tablet in a bright learning environmentReal photo
TFG · IRB / UdL · In progress
03IN PROGRESS

NECODYS – Early Dyslexia Detection

Clinical mobile app for early detection of dyslexia risk in children aged 3–6 through oral discourse analysis. Flutter app + FastAPI backend + NLP morphosyntactic analysis + unsupervised ML. Research project in collaboration with the IRB.

FlutterFastAPIPythonNLPMLPostgreSQLFirebase
Endurance athlete preparing to run on a trackReal photo
Personal · 2025
04

NeuroSport

Injury risk and overtraining prediction app for endurance athletes. Android app (Kotlin + Jetpack Compose) with offline-first architecture + FastAPI + LSTM model (RMSE 0.05) + Ollama LLM for personalised recommendations via SSH tunnel.

KotlinJetpack ComposeFastAPILSTMKerasFirebaseTFLite
Children collaborating on a robotics project in a classroomReal photo
Personal · 2024
05

FLL Scoreboard

Gamified leaderboard for First LEGO League teams integrated with GitHub. Tracks points by completed issues and PRs, betting system with virtual 'Bricks' currency, score evolution charts, and weekly statistics.

PythonJavaScriptGitHub APIGitHub Actions

Education

2022 – 2026 (in progress)

Bachelor's Degree in Computer Engineering

Universitat de Lleida

Notable grades

Artificial Intelligence10 – Honours
Software Architecture9.8 – Honours
Automata, Languages & Grammars9.7 – Honours
Advanced DB & Software Eng.9.1 – Honours
Software Engineering9.0 – Honours

Certifications

Cambridge B2 First Certificate

English · B2

AWS Cloud Foundations

Amazon Web Services

Languages

CatalanNative
SpanishNative
EnglishB2 Advanced

Skills

Languages

PythonJavaKotlinDartCPHPJavaScriptTypeScriptHTML/CSSBash

AI / Machine Learning

Scikit-learnKeras / TensorFlowXGBoostLightGBMCatBoostSHAPOptunaLSTMAutoencoders / U-NetNLPOpenCV

Mobile

Android - Jetpack ComposeFlutterTFLite

Backend & APIs

FastAPIFlaskJava SpringREST APIsNext.jsReact

Databases

PostgreSQLFirebaseSupabaseSQLite / Room

Infrastructure & Tools

AWSDockerGit / GitHubGitHub ActionsJupyterNgrok

Contact

Have a project in mind or want to chat? I'd love to hear from you.

I usually respond within 24 hours.