Data Scientist & AI Engineer

Serena Mendanha Dublin, Ireland

2× IEEE Published · MSc Data Science @ Trinity College Dublin · First-Class Distinction (1:1) · Building ML that ships to production.

scroll to explore
import sklearn IEEE ✓ 98.6% acc 1:1 🎓 model.fit() TCD 🍀
1:1First-Class · TCD
2×IEEE Published
98.6%model accuracy
Python✦Machine Learning✦IEEE Research✦NLP✦Deep Learning✦Tableau✦Data Science✦Flask✦Scikit-learn✦Healthcare AI✦ Python✦Machine Learning✦IEEE Research✦NLP✦Deep Learning✦Tableau✦Data Science✦Flask✦Scikit-learn✦Healthcare AI✦
01 // about me

Building AI
that matters

I'm a Data Scientist and ML Researcher completing my MSc at Trinity College Dublin — achieving a First-Class Distinction in Semester 1. My research is peer-reviewed and published in IEEE Xplore across two conferences.

From Raspberry Pi deployments to deep learning on clinical data, I build end-to-end — data preprocessing to production API. I'm actively seeking graduate roles in Data Science, ML Engineering & Analytics across Ireland and Europe.

2×
IEEE Papers
1:1
Distinction
7+
ML Projects
3×
Internships
02 // tech stack

What I Know

ML & AI
// core competency
Scikit-learnXGBoostRandom ForestLSTMAutoencodersNLPFeature Engineering
Data Science
// analysis & insight
PandasNumPyEDAStatistical AnalysisModel EvaluationData Preprocessing
Visualisation
// storytelling with data
TableauMatplotlibSeabornPlotlyDashboard Design
Engineering
// build & deploy
PythonSQLFlaskREST APIsGitRaspberry Pi
03 // projects

What I've Built

All on GitHub ↗
01
🛡️
Security · ML · Production Deploy

Anti-AI CAPTCHA System

Multi-layer CAPTCHA for Irish event ticketing combining video, behavioural signals & escalating illusion challenges. Built & stress-tested 5 AI bots. Deployed on Raspberry Pi.

99% → 60% AI bypass · >95% human success
PythonFlaskOpenGLRaspberry PiHTML/CSS
GitHub →
02
🔍
NLP · Deep Learning

Fake News Detector

LSTM classifier + Flask REST API. Real-time inference on user text.

LSTMFlaskNLP
GitHub →
03
🧠
Healthcare AI

Stroke Prediction

Standard, Denoising & Variational Autoencoders on clinical patient data.

KerasAutoencoders
GitHub →
04
❤️
Healthcare AI · ML

Heart Attack Analysis

Key cardiac risk factor analysis. Comparative classification with feature importance.

Scikit-learnPandas
GitHub →
05
💧
Predictive ML

Water Quality Model

pH, turbidity & TDS parameters. Clustering-based classification.

+25% accuracy
Scikit-learnClustering
GitHub →
06
📄
NLP · Pipeline

Resume Analysis Engine

ATS-powered skill extraction & job matching system.

+20% matching accuracy
NLPMLFlask
GitHub →
04 // tableau

Live Dashboards

Full Profile ↗
loading dashboard...
Sales Dashboard
Revenue trends · Product performance · Regional breakdown
Open in Tableau ↗

Want to see more? All dashboards are live on my Tableau profile.

View All on Tableau ↗
05 // research

IEEE Publications

2026
IEEE GIEST · Feb 2026

ML vs. Deep Learning for Early Detection of Neurological Disease Progression

98.6% accuracy · Precision & Recall >99%

Ensemble stacking model outperforming standalone ML and DL on real clinical patient data. Benchmarked 10+ architectures.

DOI: 10.1109/GIEST66547.2025.11387676 ↗
2025
IEEE ICSADL · Nepal · Mar 2025

ML Algorithms for Conversion Predictors of CIS to Multiple Sclerosis

Random Forest — top predictor

Comparative study: SVM, KNN, Decision Trees, Random Forest, XGBoost on clinical & MRI data for early MS diagnosis support.

DOI: 10.1109/ICSADL10933208 ↗
06 // let's talk

Open to
Opportunities

Seeking graduate roles in Data Science, ML Engineering & Data Analytics across Ireland & Europe. If you're building something with data — I'd love to be part of it.

mendanhs@tcd.ie →