01 / MACHINE LEARNING
Models, features,
and validation.
I build and evaluate machine learning models, from preparing data and designing features to testing how well they generalize.
Explore the modeling projectsENGINEERING / MACHINE LEARNING / QUANT FINANCE
stefanie.
I’m a Lebanese engineer studying data science at École Polytechnique × HEC Paris. I build machine learning systems and study how they behave on messy, real data.
Next: quantitative research at Millennium · spring 2027
HELLO
FROM PARIS
FROM PARIS
STEFANIE SAMAHA
BEIRUT → PARIS
✳
X-HEC / M1
✳
ML ENGINEERING / QUANT RESEARCH
✳
EN / FR / AR / ES
01 / a little introduction
what i work on.
My work sits between statistical learning, software engineering, and financial data.
Models, features,
02 / AI ENGINEERING
Agents &
workflow automation.
I work on AI agents and workflow automation, connecting models with tools and data to reduce manual work. In industry, that meant building and testing applied AI systems, with n8n and LangChain among the tools.
The engineering experience03 / STATISTICAL MODELING
Statistical &
numerical methods.
Binomial lattices, risk-neutral option pricing, and convergence checks against Black–Scholes. I check how pricing error changes as the lattice resolution increases.
Read the pricing project04 / QUANTITATIVE RESEARCH
From a signal
to a portfolio.
Cross-sectional equity ranking, momentum portfolios, and volatility targeting. I test the resulting portfolios with exposure constraints and drawdown diagnostics.
Explore the researchA little background
I studied Computer & Communications Engineering at the American University of Beirut, graduating with High Distinction and a 4.0 GPA. I’m now in the first year of the Data Science & AI for Business MSc&T at École Polytechnique and HEC Paris.
I enjoy the whole research process: turning a question into an experiment, checking the assumptions, and understanding why a model succeeds or fails. Financial data gives me plenty of difficult questions to work on.
A couple of lines I keep coming back to: “This is the Way.” The Mandalorian and “Do or do not. There is no try.” Yoda
02 / experience & education
the path so far.
Millennium Management
Incoming Quantitative Research Intern · Monaco
Joining a Delta One pod to work on quantitative research, datasets, and research tooling.
École Polytechnique × HEC Paris
MSc&T · Data Science & AI for Business
Currently in M1, studying statistical learning, machine learning, and mathematics. Admitted with a scholarship.
American University of Beirut
B.E. · Computer & Communications Engineering
Graduated in Computer & Communications Engineering with High Distinction and a 4.0/4.0 GPA. I also spent a lot of time tutoring engineering students.
AI engineering in industry
Agents & automation
Built and tested applied AI systems, agent workflows, and automation prototypes with tools including n8n and LangChain.
GradeLift
Co-founder · Education
Our team has helped 250+ engineering and CS students so far, through tutoring and the learning platform we’re building.
03 / selected research
research, in code.
Selected projects in machine learning, natural language processing, numerical methods, and financial data.
01 / DATA CHALLENGE
QRT Stock Prediction
Predicting next-day residual equity returns with leakage-aware features, cross-sectional ranks, and an ensemble of models.
51.54%
public leaderboard accuracy · 51.31% benchmark
View code & research notes ↗
02 / SYSTEMATIC RESEARCH
ML Alpha Ranker
From momentum and volatility features to cross-sectional predictions, portfolio construction, information coefficients, and a reproducible backtest.
IC
rank-based validation & portfolio analysis
View code & research notes ↗
03 / DERIVATIVES
Option Pricing Lattice
European and American option pricing with a risk-neutral binomial lattice, backward induction, and Black–Scholes convergence checks.
$0.001
reported lattice pricing error · N = 200
View code & research notes ↗
04 / QUANTITATIVE METHODS
Cross-Sectional Momentum
A momentum portfolio with volatility targeting, exposure constraints, and drawdown diagnostics across historical US equities.
1.31
reported long-only Sharpe · historical backtest
View code & research notes ↗
05 / NATURAL LANGUAGE PROCESSING
Emotionally Aware Digital Diary
A collaborative NLP project exploring how a diary can respond to what you write. Fine-tuned DistilBERT models extract emotions, personality traits, and events, which feed a GPT-2 response generator.
92.6%reported emotion classification accuracy
Read the project paper ↗
04 / small experiment
a little
probability.
A small random-walk experiment: change the bias, run a few paths, and compare the average with E[Sₙ] = n(2p − 1).
simulation ready
05 / beyond code
a few other obsessions.
01 / CONTROLLER IN HAND
✦
video games
Story-heavy RPGs, great worlds, shooters, and racing games. If the writing, soundtrack, or mechanics are good, I'm probably interested.
Persona 5 Royal · Baldur’s Gate 3 · Cyberpunk 2077
see the games I've been playing +
Persona 5 Royal
Cyberpunk 2077
Baldur’s Gate 3
Red Dead Redemption 2
Titanfall 2
Battlefield 6
Metaphor: ReFantazio
Final Fantasy XV
Ghost of Tsushima
Marvel’s Spider-Man 2
The Last of Us Part I
The Last of Us Part II
Psychonauts 2
Forza Horizon 6
gym
Gym sessions and fitness classes. Staying fit is one of the things I find most fulfilling.
02 / CURRENTLY ON THE SHELF
❝
books
Existentialism, Russian classics, French literature, and ambitious fantasy. I enjoy a book that stays with me after I put it down.
Sartre · Kundera · Dostoevsky · Sanderson
open my bookshelf +
La Nausée: Jean-Paul Sartre
The Sword of Kaigen: M. L. Wang
Mistborn: Brandon Sanderson
Anna Karenina: Leo Tolstoy
L’Insoutenable Légèreté de l’être: Milan Kundera
Le Livre du rire et de l’oubli: Milan Kundera
La Plaisanterie: Milan Kundera
L’Étranger: Albert Camus
Hamlet: William Shakespeare
The Brothers Karamazov: Fyodor Dostoevsky
The Priory of the Orange Tree: Samantha Shannon
Les Liaisons dangereuses: Pierre Choderlos de Laclos
Les Fleurs du mal: Charles Baudelaire
LEGO
Building things, just for fun.
03 / ALWAYS PLAYING
♫
reggaeton
My taste is varied, but reggaeton dominates.
See what’s on repeat ↗languages
Arabic, English, French, Spanish. Currently learning Chinese.
Rubik’s cubes
More than 20 puzzles, in different sizes and shapes.
✈
cities & long walks
Travel, new places, art, photography, and finding the good cafés.
✎
teaching & campus life
Four years of tutoring engineering students, plus collaborative projects and campus experiences at AUB.
A FEW PHOTOS FROM MY CAMERA ROLL
a few favorite places & moments
06 / a few more tabs
the other pages.
Travel photos, favorite albums, a few milestones, and some things about me that did not fit on a CV.
07 / say hello
say hello.
Have a research question, a project in mind, or a book I should read? I’d love to hear from you. I’m always happy to talk ML, compare favorite games, or meet someone new.
stefanie.samaha@polytechnique.eduOr browse the code on GitHub.