AI Engineer
Pubblicato il 24-10-2026 - AI Lynx in Napoli
In a financial time series, the signal—when there is one is weak. Most experiments lead nowhere, and a single validation error produces results that look beautiful and are false. The job is to find what holds up out of sample and take it into production, inside an app that’s about to launch.
We’re AI Lynx, based in Naples. We trained, validated, and deployed the model that runs inside VEGA ourselves—VEGA is our quantitative investment analysis app for self-directed investors, and it’s ready to go to market. We carry out our research together with universities and research centres: that’s how we work. More fintech products are in the pipeline, the next one for banks and investment managers, and we’ll build their models with you.
You’ll work daily with our CTO, on both the experiments and the code. The model you validate goes into production inside VEGA: at this stage there’s no other team in between.
What you’ll do
- Design and run reproducible experiments on financial time series: features, models, validation, and historical simulations.
- Test your results out of sample. House rule: a result that doesn’t replicate doesn’t count.
- Take models into production, inside VEGA and the products in our pipeline.
- Work with researchers from the universities we collaborate with.
- Use frontier AI tools, including for research, and work with us to refine how we use them.
What to expect
- A gross annual salary of €28,000–€35,000, plus benefits.
- Experiments that are yours, end to end: from hypothesis to model in production.
- Most experiments never make it into production. If you need to see a result every week, you’ll struggle here.
- A small team, with our CTO as your day-to-day point of reference.
- You can publish, within one limit: general scientific work, yes; details of the financial domain, no.
- Remote working from Italy, with a few meetings in Naples during the year, or on-site at the Centro Direzionale. Start date: autumn 2026.
What we’re looking for
- A degree in computer science, engineering, mathematics, or physics, or an equivalent background you can demonstrate.
- Grounding in ML/DL fundamentals—training, evaluation, generalisation—and solid mathematics: probability, statistics, and linear algebra.
- At least one substantial ML project, academic projects included, whose choices, mistakes, and limitations you can discuss.
- Tracked, reproducible experiments, and the patience to stay with a problem that won’t give way for weeks.
A plus: a project outside your main area (CV, NLP, or something else), which tells us how quickly you get up to speed in a new domain; experience with financial data or an interest in finance and trading; publications or competitions.
We’ll respond within 10 business days of your application. Interviews begin in October, and the process from the first interview to an offer takes 3–4 weeks.