Offer Description

Job Opening for ML/AI R&D;

(Spanish Startup - Spanish PhD or officially certified in Spain required due to be funded by European ERDF funds)

Two AI experts - Sedna International SL (startup with a Neotec project)

Location

Vigo, Spain. On-site/telematic flexible with adaptable hours also.

Start date

Between October / November / December 2026.

Duration

Until June 2028.

Contract / Pay

44,700 € per year (plus possible variable depending on results). Category 1 (Spanish Social Security highest contribution group).

Candidate profile: Someone who is passionate about AI/ML and R&D.; In this case it will be applied to financial time-series data (quantitative trading strategies). We look for research skills and drive; a background in finance is not a requirement. Your degree field doesn't matter -AI, mathematics, computer science, physics, engineering- as long as you are comfortable with AI/ML techniques and time series, and love what you do.

About us

Sedna International is a fintech startup created for developing artificial intelligence for ethical and sustainable investing. We hold the Innovative SME Seal from the Ministry of Science and Innovation, an Activa Startups pilot project from the Ministry of Industry, the CDTI Neotec Seal, and we are now launching the Neotec project. This project is funded by the Galician Innovation Agency (GAIN) and runs until 30 June 2028. To carry it out, two R&D; (Research & Development) positions are offered, one of them reserved for a PhD holder.

Behind Sedna is a founding team with more than twenty years leading business and R&D; projects in demanding environments: from instrumentation for high-energy physics (particle accelerators) and astrophysics, to supercomputing -including participation in a €33 million project-. The Activa Startups pilot project was already in the AI/Fintech area.

The project

We will build an AI-first multimodal predictive system. It will be based from historical and real-time data, not from economic theories or market models -AI applied directly to the data.

The goal is to generate Green Alpha,radical economic growth by investing solely in ESG assets (Environmental, Social and Governance, i.e. sustainable and ethical investing) and, in the future, to bring it within reach of the retail investor, transforming the ordinary citizen's access to responsible investing built on cutting-edge technology.

What you'll work on

T he work is R&D; into quantitative investment strategies (Quant Trading): designing, backtesting, implementing and validating systematic strategies. A good starting point for understanding the kind of real strategies you'll work on is strategy library. There will be a familiarisation phase with the jargon, the fundamental structure of the market and various indicators; a second phase researching known strategies; and a third phase modifying and creating our own.

In our project, the investment universe focuses on ESG assets (sustainable and ethical). The goal is to generate Green Alpha -economic growth by investing solely within that universe- which narrows the search space and turns the problem into a more specific and, technically, more interesting one, as well as a more ethically responsible one.

In practice you'll move constantly between these tasks:

- State-of-the-art monitoring. Read papers, forums and repositories; follow what's being published in Quant Trading and in machine learning applied to financial series,



and spot what's worth trying.
- Analysis of existing strategies. Study published strategies, understand why they work (or why they worked), reproduce them and use them as a starting point or as a control.
- Experiment design and rigorous backtesting. Clean backtests: point-in-time data, no survivorship or look-ahead bias, with realistic costs and slippage. Understand backtesting techniques scrupulously: overfitting is one of the main enemies of this work.
- Out-of-sample validation and forward testing. Honestly separate what merely fits from what generalises, and take whatever survives towards real-time validation.
- Working with time-series data. Acquisition, cleaning and transformation of historical and real-time series, feature generation and synthetic data to train agents.
- AI/ML modelling. Time-series models, reinforcement learning and multimodal architectures applied directly to the data, without starting from prior economic theories.
- Relentless iteration. Read the results, decide the next step, discard what doesn't work and start again -many times! Build agents that systematise comparison or competition between strategies, etc.

The profile we're looking for

We're not specifically looking for quants. If you are one, great. If not, you don't need to come from finance or have trading experience: understanding technical indicators and market concepts is a matter of weeks; you'll pick it up quickly.

The difficulty lies in R&D; skills: knowing how to find your way through fast-moving literature, designing an experiment that truly answers the question you're asking, distrusting your own results, handling time-series data with ease, and having the statistical judgement to tell signal from noise. The good part is that you can come from fields such as mathematics and statistics as much as from computer science and AI, physics, engineering or others.

To that we add a second condition: genuine enthusiasm for what's happening right now in computing -AI, machine learning, coding agents, multi-agent systems, etc.

We want people whose curiosity is sparked by the new tools and the historic technological moment we're living through. People already using the latest tools, who understand that the way we do research has changed: today, with good agentic tools, you can cover in an afternoon what used to be weeks of work -if you have enough knowledge and judgement.

Thirdly, and perhaps the most important: a real hunger to take on a genuine challenge. This is R&D; at a startup in its earliest phase, with enormous potential impact. We are looking for someone who understands its value, who understands the effort behind being able to offer an opportunity like this, and how unique the occasion is. There will be a thousand failed attempts for every result that counts, frustrating days and euphoric days. In return, an intense experience, with real freedom and flexibility, where everything you discover leaves a mark.

On technical knowledge

Given the nature of the project,



learning will be continuous for the candidate; but in order to have the ability to learn and the judgement to go with it, we do value prior knowledge and experience.

- Python programming. Even though writing code now tends towards zero, having programming experience, in structuring code, is necessary to grasp everything the work involves. Specific experience with basic tools such as Pandas, NumPy, SciPy, matplotlib is always welcome.
- Typical ML/AI frameworks: scikit-learn, TensorFlow, PyTorch. Understanding of the methods, not just use of libraries. Experience in evaluation, validation and hyperparameter optimisation.
- Advanced time series: TFT, N-BEATS, DeepAR, GNNs, hybrid architectures that integrate LLMs with time-series models.
- Reinforcement Learning: Actor-Critic (PPO, SAC, A2C, DDPG), Value-Based (DQN, QR-DQN), multi-agent. Stable Baselines3, RLlib or TF-Agents.
- Data acquisition: setting up ingestion from APIs and external sources -now largely delegable to agents-; what we value is the judgement to design it, validate it and keep it reliable.
- Unstructured data and NLP: extracting information from reports and news; familiarity with LLMs and agent-based systems.
- Research skills: autonomy to design experiments, read their results and decide the next step.
- Teamwork: The idea is for the two researchers to work closely together to face the challenges. We'll create the working environment and organisation that makes the most sense, one that makes the experience as rewarding as possible. It will be a time to remember.

We remark that we will value candidates who have the passion and the drive to take on this challenge more than whether they've had more or less experience with the technologies listed above. If this is your world, what you truly like and are motivated by, here we offer the time, resources and goals to develop your talent to the fullest.

What we offer

The freedom of a tech startup with founders who have spent two decades doing top-level R&D;, participation in a Neotec project and a working culture built on passion for what we do, on enjoying it, on pushing the limits. A project at an exciting stage, where every contribution leaves a mark and the possibilities are wide open and the sky is the limit. Genuinely hard problems, with moments of frustration and euphoria. An initial contract until June 2028, at Category 1 from day one.

If all of this appeals to you, please write to us at with your CV and, if you like, anything you'd care to tell us about yourself. Web:

Note for PhD holders: the doctorate must be completed before the hiring date; if it is from a non-Spanish university, it must be officially validated (homologado) in Spain before the hiring date to comply with the terms. Even so, a PhD holder without Spanish recognition could always apply for the graduate position instead.

Where to apply E-mail

[email protected]

Requirements

Research Field All Education Level PhD or equivalent

Languages ENGLISH Level Excellent

Research Field Computer science » OtherMathematics » Computational mathematics

Additional Information

Website for additional job details

Work Location(s)

Number of offers available 2 Company/Institute Sedna International Institute SL Country Spain State/Province Pontevedra City Mos Postal Code 36417 Street Louriñó-Sobráns Geofield

Contact State/Province

Pontevedra City

Mos Website

Street

Louriñó-Sobráns Postal Code

36417

STATUS: EXPIRED

Share this page

📌 ML/AI Researcher - R&D in Fintech Startup (Vigo)
🏢 Reconocida empresa
📍 Vigo