Hello, I'm
Francisco Costa

Francisco Costa

MSc Business Analytics  ·  BSc Aerospace Engineering  ·  Data Analytics & ML

📍 Lisbon, Portugal

View My Work Get In Touch
Scroll

About Me

I'm a former high-performance judo athlete. A sport that shaped how I approach every challenge: with discipline, adaptability, and a drive to keep improving. That same mindset led me from a Bachelor's in Applied Mathematics, through a Bachelor's in Aerospace Engineering at Instituto Superior TΓ©cnico, and a year as a trainee at Lisbon Airport, to a Master's in Business Analytics at Nova SBE.

I'm drawn to complex systems and the kind of problems where data can genuinely change how decisions are made. Whether that's forecasting patient demand in a resource-limited hospital, detecting solar panels from satellite imagery, or building models that hold up under real conditions.

Whether it's cleaning the messy data, building ETL pipelines, to training machine learning models and designing analytical frameworks, I care about solutions that work beyond the notebook. Currently looking to grow in data analytics, data science, or data engineering, in environments that value rigour, ambition, and curiosity.

Francisco Costa competing in judo

Skills & Technologies

🤖
Machine Learning
Scikit-learnXGBoostLightGBMRandom ForestLogistic RegressionK-MeansProphet
🧠
Deep Learning
PyTorchHuggingFaceRoBERTaViTLSTMYOLOv8FastTextGemini API
📊
Data Analysis
PythonPandasNumPyStatsmodelsSciPyRJupyter
🗃
Data Engineering
SQLMySQLMariaDBETL PipelinesApache SparkRAPIDS cuDFGit
📈
Visualisation & BI
TableauPower BIStreamlitMatplotlibSeabornPlotly
🛠
Engineering & Other
Matlab / SimulinkSolidWorksRISC-V / AssemblyDbSchemaMS Office

Experience & Education

Experience
JUNE 2025 — JAN. 2026
Data Analyst & IT Infrastructure Intern
Change ETS (NGO) · Ampefy, Madagascar
Master's thesis internship. Built ETL pipelines and restructured the hospital's legacy databases (MariaDB). Developed predictive models for patient volume forecasting and automated dashboards. Strengthened IT infrastructure and worked directly with hospital leadership. 2 months on-site in rural Madagascar.
OCT. 2023 — SEPT. 2024
Trainee — VINCI Internship Program
ANA Airports · Lisbon, Portugal
Contributed to the development and rollout of APMS (Airport Pavement Management System) across 10 airports managed by VINCI/ANA. Led training sessions for engineering teams. Also supported electronic systems maintenance at Lisbon Airport and drafted technical specifications for system upgrades.
2016 — 2022
High-Performance Judo Athlete
Sporting Clube de Portugal
Competed at the highest national level (Level B & C). Two-time Teams Champions League champion. Represented Portugal at the World Judo Championship Juniors 2019, 5+ European Championships, and 15+ international events.
Education
SEPT. 2024 — JAN. 2026
MSc Business Analytics — 18/20
Nova School of Business & Economics · Lisbon
Specialisation in Machine Learning and AI.
Thesis grade: 19/20 — "Operational Intelligence in Humanitarian Healthcare" applied to Saint Paul Hospital, Madagascar.
Coursework: Advanced ML, AI Impact on Business, Data Visualisation, Algorithmic Governance, Blockchain.
OCT. 2020 — JULY 2023
BSc Aerospace Engineering — 14/20
Instituto Superior Técnico · Lisbon
Rigorous engineering foundation in mathematics, physics, and systems thinking.
Coursework: Flight Stability, Flight Control, Flight Testing, Aerodynamics, Orbital Mechanics.
SEPT. 2017 — JULY 2020
BSc Applied Mathematics — 12/20
Faculdade de Ciências, Universidade de Lisboa · Lisbon
Strong quantitative foundation.
Coursework: Calculus I–III, Linear Algebra, Numerical Analysis, Mathematical Models, Programming I & II.

Selected Projects

🏥
Master's Thesis · Field Work
Operational Intelligence at Saint Paul Hospital, Madagascar
ETL pipeline + predictive analytics for a humanitarian hospital in rural Madagascar. Laboratory automation tripled capacity; 192k records analysed.
PythonMySQLProphetStreamlitETL
Group Project · Computer Vision
HuggingSUN — Solar Panel Detection App
YOLOv8 segmentation model detects solar panels from satellite imagery. Full Streamlit app with area estimation, cost calculator & nearby services locator.
YOLOv8PyTorchStreamlitGoogle Maps API
🧠
Individual Assignment · Deep Learning
Multimodal AI-Generated Fake Review Detection
Compared CNNs, LSTMs, RoBERTa, ViT, and Gemini for detecting AI-generated reviews. RoBERTa fine-tuned achieved 99% accuracy.
RoBERTaViTLSTMGeminiPyTorch
🌿
Group Project · Time Series ML
Air Pollution Forecasting — Beijing
Random Forest classifier forecasting hourly PM2.5 levels for Beijing. 81.7% accuracy with 87.9% recall on high pollution class.
Random ForestXGBoostFeature EngineeringScikit-learn
🛒
Individual Assignments · ML
Telecom Churn & SmartRetail Intelligence
Churn prediction with LightGBM (96.5% accuracy) + customer segmentation, loyalty prediction, and recommendation system.
LightGBMK-MeansRec. SystemsScikit-learn
📊
Group Project · Statistical Analysis
Employee Attrition: Age & Work-Life Balance
Logistic regression & t-test on IBM HR dataset. CEO claimed 15% churn from intl. plan holders β€” data showed 27.59%.
PythonStatsmodelsLogistic RegressionT-Test
Get In Touch

Let's Connect

Open to new opportunities, interesting projects, and conversations about data, impact, and everything in between.