Press on.
Keep working through the part that does not come easily.
enode — a student-owned technical association at ESADE
A student-owned association committed to excellence in data science, quantitative finance, and applied AI — building rigor into work that actually ships.
01research / ai
02quant
03ai4law
SCROLL OR CLICK TO EXPLORE
↓
01 / The associationESADE · Since 2024
Learn the foundations.
Make something
of them.
We are a student-owned collective working across data science, quantitative finance, and applied AI.
Workshops to understand the mathematics. Projects to put it to work. Conversations with people who have gone further.
Meet enodeSant Cugat & Pedralbes, Barcelona
02 / Three departmentsA shared standard
Find your department
01research. The mathematics beneath the model.Foundations · Architectures · Systems 02quant. From uncertainty to understanding.Mathematics · Strategies · Engineering 03ai4law. The institutions our systems touch.Law · Technology · Accountability03 / Selected workThe work is the evidence
A record of
learning by doing.
Projects, workshops, and contributions.
Some finished. Some still taking shape.
04 / ProgrammeIn the room, together
Keep learning.
Workshops, reading groups,
and visiting perspectives.
Explore the event archive
| Mon | Tue | Wed | Thu | Fri | Sat | Sun |
|---|
Bring a question.
Stay for the work.
Engagement, initiative, and the willingness to keep going.
That is where membership begins.
The association01 / About
Student-owned.
Work-led.
An academy, a lab, and a place to think together. Founded at ESADE in 2024, built by the people who show up.
2024Founded
~40Members across cohorts
30+Workshops since launch
02Campuses
Our starting pointUnderstanding
Understanding
takes practice.
enode is a student-owned technical association at ESADE, operating across Sant Cugat and Pedralbes. Research and Quant are based in Sant Cugat; AI4Law brings the legal perspective from Pedralbes.
We run workshops and learning sessions, build applied projects and research notebooks, and bring external speakers, practitioners, and alumni into the room through localhost.
A small board coordinates the association. Members make the workshops, projects, and conversations happen.
Three principles
Show your work.
Mathematically describe what you implement. Implement what you describe.
Make it real.
A notebook, a workshop, a paper, a strategy in the paper-trading account. The artifact is the proof.
An ongoing recordSmall beginnings.
Small beginnings.
Concrete work.
- 2024 / Q4
Founding cohort
First members, workshop calendar, and the launch of Quant.
- 2025 / Q1
Natural-language queries · Barcelonactua
A database querying tool for an NGO. The engagement is now paused.
- 2025 / Q2
Agentic prototype · Supersonik
A lead-production demo, subsequently archived as the association refocused on fundamentals.
- 2025 / Q3
GPT-2 & Llama, from scratch
Members rebuilt both architectures end-to-end.
- 2025 / Q4
AWS data platform
Terraform-managed ETL and paper-trading on Alpaca.
- 2026 / Q1
AI4Law launches
A Pedralbes department and a shared cross-campus programme.
- 2026 / Q2
Backtesting & research contributions
Backtesting engine v0.4 and member contributions to a Spanish research group.
Run by membersThe people
The people
behind the work.
Board appointments are in an ongoing election process. The structure below describes how the association is organised.
Board & operating structure
- President & executive board
- QuantDepartment head · Workshop lead · Projects lead · Associates
- Research / AIDepartment head · AI Foundations lead · Applied ML lead · Associates
- AI4LawDepartment head · Content / Doctrine lead · Associates
- Marketing & CommunityDepartment head · Technical and legal content leads · Associates
- OperationsDepartment head · Events coordinator
01 / Sant Cugat
Research & Quant
ESADE Sant Cugat campus.
Workshops in Aula 001 and 005.
02 / Pedralbes
enode@esade.edu AI4Law
ESADE Pedralbes campus.
A shared association, across disciplines.
Find the work you want to be part of.Join enode
enode / 01Sant Cugat
research.
The how and why beneath modern computing. Mathematical foundations, the mechanics of code, and projects that make understanding visible.
Field note 01 / OptimizationUnderstanding
Understanding
the way down.
A loss landscape. A starting point. A sequence of decisions. Change the learning rate and watch the same gradient lead to a different journey.
The Rosenbrock landscape has a curved valley leading to a minimum at (1, 1). The controls update a real gradient-descent trajectory.
ROSENBROCK / GRADIENT DESCENTLoss 24.2000 · Step 0000
Small steps are patient. Larger steps can overshoot the valley. An illustrative loss function, with the actual gradient underneath.
From theory to systemsBeyond using
Beyond using
the library.
We want to understand what our tools do, why they work, and where they fail. Our long-term aim is a collective repository of workshops and projects that prepares members for technical work and further study.
The mathematics
Maximum likelihood, change of variables, transformation of densities, optimization, and automatic differentiation.
The model
Embeddings, transformers, mixture density networks, diffusion, graph neural networks, and applied ML.
The implementation
Memory, performance, and lower-level code. Learning to read C++ and Rust, and to understand what sits beneath a training loop.
The conversation
If you have mastered a library or built a side project, bring it into the room. Teaching is another way to test understanding.
Work to dateBuilt, tested,
Built, tested,
learned from.
GPT-2 & Llama, from scratch
A member group replicated both architectures end-to-end. Rebuilding a system gives its abstractions something concrete to stand on.
Learning alongside researchers
Select members have contributed to work with the Spanish research group ICSO, including a paper on forecasting limits in residential water demand, described in our project record as under review.
Applied work, a considered pause
A natural-language database query tool for Barcelonactua and an agentic lead-production demo for Supersonik. Consulting is currently on the back-burner as we focus on fundamentals.
Next in the room.
Programme & archiveCuriosity, with follow-throughYou do not need to
Join ResearchYou do not need to
be a researcher.
Bring statistics, calculus, linear algebra, and Python—or a clear commitment to getting there. Engagement, initiative, and consistency matter. Systems curiosity is a welcome bonus.
enode / 02Sant Cugat
quant.
Understand the mathematics. Write the strategy. Build the infrastructure that makes the work reproducible.
Field note 02 / UncertaintyOne beginning.
One beginning.
Many outcomes.
A path tells one story. A distribution tells you what that story leaves out. Adjust volatility to see uncertainty take shape.
A geometric Brownian motion simulation produces a range of possible outcomes. Increasing volatility broadens their distribution.
128 SIMULATED PATHS / 1 YEARStarting value 100
Geometric Brownian motion · 5% annual drift · 252 steps. Shaded area: sample 5th–95th percentiles. Illustrative simulation.
01 / EducationThe mathematics
The mathematics
comes first.
Education is a core part of the department. We want to describe our strategies mathematically and understand the assumptions we turn into code.
Our workshops move from real analysis and optimization into stochastic processes and financial applications. We document the teaching so future members can follow the theory and reuse the material.
From theory to labs02 / Research & strategiesChoose a question.
Choose a question.
Work it through.
Small pods choose the mathematical trading questions they want to pursue, then bring the results back through write-ups, reviews, and clean implementations. We run strategies in a paper-trading account; a small student-led fund remains a longer-term ambition.
- 01
Options & Monte Carlo
Pricing intuition, volatility, sensitivities, simulation-based pricing, and risk estimation.
- 02
Returns & dependence
Distributions, tails, correlation, regime shifts, and the limits of diversification.
- 03
Portfolio construction
Black–Litterman, optimization, momentum, pairs trading, and volatility forecasting.
03 / EngineeringReproducibility is part of the work
A strategy needs
somewhere to run.
IngestSource dataPrepareClean & featureEvaluateBacktestDeployPaper-trade
Data infrastructure
ETL and storage on AWS, managed with Terraform. A repeatable path from source data to usable features.
Research infrastructure
A consistent backtesting engine, versioned datasets, proper packaging, and environments that travel beyond one laptop.
Next in the room.
Programme & archiveTwo profiles. Shared work.Model the idea.
Join QuantModel the idea.
Or build its foundations.
Math, statistics, and Python on the research side. Databases, Docker, cloud infrastructure, and reliable systems on the engineering side. Either profile is welcome; the best teams have both.
← enode / quantWorkshop note
From theory
to labs.
Before writing trading code, understand the mathematics, where the assumptions enter, and how a model becomes an implementation.
Slow down, in the right way.
The quant track is not meant to be a collection of disconnected finance demos. We want members to move from definitions to proofs, from proofs to models, and from models to code that can be inspected, defended, and improved.
The first block built the mathematical base: real analysis, numerical thinking, optimization, and the discipline of stating assumptions precisely. The second moved into stochastic processes and financial applications, where the same theory becomes simulation, pricing intuition, and hands-on labs.
The point was not to cover everything. It was to give members enough structure to read future material seriously, ask better questions, and build strategies without treating the mathematics as decoration.
Who guided the work.
Marius Oltean
Marius supported the mathematical foundations: real analysis, optimization, and the bridge from abstract concepts to quantitative reasoning. At ESADE, he coordinates mathematics courses in the BBA and BAIB programmes. His background is in mathematical physics, specialising in general relativity and numerical analysis.
Adrià Garcés Ortiz
Adrià led the stochastic workshops, final financial applications, and hands-on labs. He is a PhD student with an MSc in Physics of Complex Systems & Biophysics and a BSc in Physics from Universitat de Barcelona.
We are grateful to both for helping a young department bring rigor, curiosity, and implementation together.
Follow the material.
Instructor content and slides are available in the public Drive; stochastic-process implementations are in the GitHub repository.
enode / 03Pedralbes
Intelligence.
In context.
AI4Law brings legal thinking into the room with technology. For aspiring law professionals who want to understand the systems reshaping their discipline—and the institutions those systems touch.
Field note 03 / A decision, examined
Change the lens.
A technical process carries
human questions at every step.
- 01
Data
What enters - 02
Model
What interprets - 03
Decision
What follows - 04
Review
Who can question
§ 01
Whose data, and on what terms?
Trace where the information came from, what its use was intended to be, and how that use follows a person through the system.
A way to ask better questions. Explore each perspective.
The syllabus is openRead closely.
Read closely.
Think across
disciplines.
§ 01Regulatory landscape+
The EU AI Act, GDPR, sectoral compliance, and the operational shape of high-risk systems.
§ 02Doctrinal analysis+
Reading judgments and statutes for evidence, scope, and limits, with the same care technical readers bring to papers.
§ 03Applied prototypes+
Document review, contract drafting, and evidentiary tooling, built with lawyers in the room.
§ 04Ethics & data privacy+
Where machine-learning decisions meet due process, consent, and accountability.
§ 05Cross-campus seminars+
Joint sessions with Research and Quant. Lawyers explain problems; technologists explain tradeoffs.
§ 06Career & practice+
Conversations with legaltech founders, in-house counsel, judges, and academics shaping the discipline.
One associationA different lens.
A different lens.
A shared standard.
AI4Law operates under enode’s charter and editorial standard, from the Pedralbes campus. Membership is open to students of either discipline.
Research and Quant interrogate models. AI4Law examines the institutions those models touch, and what credibility means when the artifact is a contract, a brief, or a verdict.
Next in the room.
Programme & archive ↗Law meets the systems around itBring your
Join AI4LawBring your
perspective.
A place for legal scholars and technical thinkers to learn from each other, through close reading, practical work, and questions worth staying with.
Become part of enodeCuriosity is a starting point
Press on.
You do not need to arrive as a researcher or a quant. Bring engagement, initiative, and the willingness to put in the work.
