🧠 ML Engineer (RL Infra, Founding)

VaultMind — Delft, Netherlands | Hybrid

Compensation: Equity • Salary begins post-funding

Stage: Pre-seed round in progress with multiple investors and incubators

🚀 About VaultMind

VaultMind is a research-driven company developing artificial intelligence that anticipates and prevents blockchain vulnerabilities before they become exploits.

Originating within the TU Delft ecosystem, VaultMind is a deep-tech startup pushing the boundaries of applied AI for real-world security in one of the fastest-growing technological fields — blockchain.

We’re seeking exceptional talent to join us at the frontier of autonomous cybersecurity — specifically, an ML Engineer to build the RL infrastructure and training systems behind our learning engine.

If you hate corporate bureaucracy, crave real autonomy, and want to move your career 10× faster — this is your chance. Especially if you’re looking for your future AlphaGo moment in cybersecurity.

🧩 What You’ll Do

• Collaborate closely with our AI Research Lead to accelerate RL experiments and improve training stability.
• Build and maintain VaultMind’s end-to-end RL/ML training pipeline.
• Extend and refine our Gymnasium-based EVM environment for more complex contracts and observation spaces.
• Integrate advisor model outputs and analysis signals into RL state and reward mechanisms.
• Implement robust experiment runners, training loops, logging, and evaluation tools.
• Develop efficient data-loading and preprocessing utilities for traces, logs, and exploit data.
• Optimize and manage GPU training environments (RunPod/H100) for scalable, reproducible training.
• Convert research prototypes into modular, production-ready RL infrastructure supporting multi-contract learning.

🎓 Ideal Background

• Strong experience with Python and PyTorch.
• Proven ability to build ML training pipelines and experiment infrastructure.
• Familiarity with RL frameworks (SB3, Gymnasium, or custom RL systems).
• Comfortable working with GPU environments, containers, and performance optimisation.
• Experience with experiment tracking (W&B, MLflow, or equivalent).
• Solid engineering practices: clean code, modular design, reproducibility.
• (Bonus) Experience with RL algorithms (PPO, A2C, SAC) or custom implementations.
• (Bonus) Interest in security, blockchain, or smart-contract systems — not required, but helpful.
• Self-driven, comfortable working in a fast-moving research-focused environment.

💬 No startup experience required — you’ll learn directly from an experienced founder and gain hands-on exposure to real venture building.

💎 What We Offer

• Founding-level equity (vesting from day 1).
• Salary activates after funding closes.
• Work on one of the first autonomous security AIs with tangible real-world impact.
• Deep technical ownership within a focused, mission-driven environment.
• The chance to be part of multiple upcoming incubator programs (YES!Delft and other innovation initiatives).
• A rare opportunity to experience true entrepreneurship — from research to real company building — in what could become the next DeepMind for cybersecurity. (Yes – it’s got rhythm, we know.)
• Hybrid setup with Delft office as our base.

📨 How to Apply

We’re not hiring employees — we’re building an early technical founding team to invent the next generation of autonomous cybersecurity.

You can apply directly here — or email your GitHub / engineering portfolio and a short note on why RL infrastructure and ML systems excite you to [email protected]. Use Subject Line: “ML Engineer — Early VaultMind Team.”

Note for Previous Applicants

If you previously applied for the AI Research Lead role and believe your strengths align more with ML engineering and RL infrastructure, we encourage you to apply here as well.

Don’t worry about number of applicants when you apply — we’re not looking for the fastest. We’re looking for the right mind.

VaultMind — Built with discipline. Secured with intelligence.