Important: please apply through our website instead of here on LinkedIn, you will find the link to the application form in the Careers link at the top of https://www.tropicosecurity.com/

About Tropico Security

At Tropico, we’re not just building the next generation of cyber defense, we’re rewriting the rules of the game. We’re a small, fast-moving crew with zero corporate fluff, where ideas beat titles and trust beats micromanagement. Everyone owns their work here, has full autonomy to experiment (and sometimes break things), and is encouraged to challenge the status quo. We work hard, laugh often, and believe the best security tools are built by curious minds who aren’t afraid to push boundaries. If you like solving big problems without layers of bureaucracy in the way, you’ll fit right in.

Tropico builds AI-powered decoys that lure cyber attackers into revealing themselves, think of it as setting digital booby traps that collect intel before real damage is done. Our emulators mimic real IT services, apps, and devices so convincingly that intruders can’t resist, giving defenders early warning and actionable insights to stop threats.We are backed by the European Space Agency Business Incubation Center and some of the best Angel Investors in Europe!

The Role

We’re looking for an AI/ML Engineer to push the boundaries of how machine learning can predict and preempt threats. You’ll fine-tune open-source LLMs (like Mistral), develop in-house ML models for real-time intrusion analysis, and help shape AI pipelines that anticipate attacker behavior based on telemetry from our emulation layer.

This is a hands-on role where your work directly impacts how organizations defend against sophisticated adversaries.

Responsibilities

• Fine-tune and optimize large language models (LLMs) such as Mistral for adversary engagement use cases.
• Design and train custom ML models for predictive threat detection and attacker behavior profiling.
• Collaborate with security researchers to translate adversary TTPs (Tactics, Techniques, and Procedures) into machine learning features.
• Develop data processing pipelines for high-volume telemetry from IT, OT, and application emulators.
• Benchmark and evaluate AI models for accuracy, latency, and operational robustness in live environments.
• Stay ahead of advancements in AI and cybersecurity, and rapidly prototype relevant capabilities.

Requirements

• Strong proficiency in Python and deep learning frameworks (PyTorch, TensorFlow).
• Experience fine-tuning, training, and deploying large language models and transformer architectures.
• Solid grasp of machine learning fundamentals, including supervised/unsupervised learning, embeddings, and model evaluation.
• Experience with data engineering for high-throughput, real-time systems.
• Familiarity with cybersecurity concepts, intrusion detection, or adversary simulation is a strong plus.
• Ability to work in a fast-paced, mission-driven environment.
• Bonus points: in the application form, append "I read the requirements" after your last name.

Nice to Have

• Experience with MLOps tools for deploying and monitoring AI models in production.
• Familiarity with MITRE ATT&CK framework.
• Experience with synthetic data generation for model training.
• Knowledge of OT security and ICS protocols.
• Availability to travel for office meet ups in Milan and team building events.

Location

You can join us from our Milan city-center office or work remotely, whatever makes you happiest and most productive. Every quarter, we get together somewhere in Europe for team-building adventures, great food, and plenty of laughs. It’s our way of recharging, bonding, and gearing up to get back out there and keep the bad guys on their toes.