Helsing is a Munich-based defence AI and drone company that develops defence software, drones, and military systems for air, land, sea, and underwater operations.
<h2 id="Who-we-are" data-local-id="d5b97257-6724-403a-a531-87cff62cc36b" data-renderer-start-pos="1"><strong>Who we are</strong></h2> <p data-renderer-start-pos="13" data-local-id="83eb7e9a3baa">Helsing is a defence AI company. Our mission is to protect our democracies. We aim to achieve technological leadership, so that open societies can continue to make sovereign decisions and control their ethical standards.<br><br>As democracies, we believe we have a special responsibility to be thoughtful about the development and deployment of powerful technologies like AI. We take this responsibility seriously.<br><br>We are an ambitious and committed team of engineers, AI specialists and customer-facing programme managers. We are looking for mission-driven people to join our European teams – and apply their skills to solve the most complex and impactful problems. We embrace an open and transparent culture that welcomes healthy debates on the use of technology in defence, its benefits, and its ethical implications.</p> <h2 id="The-role" data-local-id="f50c60e0-a387-4c79-9800-26406946d9f2" data-renderer-start-pos="827"><strong data-renderer-mark="true">The role</strong></h2> <p data-renderer-start-pos="837" data-local-id="5ad2fe019325">At Helsing, we are pioneering the future of autonomous decision-making for defence. Our work spans the full AI landscape, including high-volume data processing, RL agent training, and large-scale foundation models. As a member of a cross-functional team, you will architect and implement the tools and platforms that enable these breakthroughs. Your focus will be on abstracting complex distributed systems to maximise training throughput and developer velocity. We are looking for engineers who can navigate the convergence of machine learning and systems engineering to build robust, scalable platforms.</p> <h2 id="What-you-will-do:" data-local-id="9a529709bd6b" data-renderer-start-pos="1444"><strong>What you will do:</strong></h2> <ul class="ak-ul" data-local-id="4a9d59c8bc07" data-indent-level="1"> <li> <p data-renderer-start-pos="1465" data-local-id="9d4c92dc0319">Extend our highly integrated deep learning frameworks (built on top of PyTorch), making them efficient and easy to use for a wide range of use cases.</p> </li> <li> <p data-renderer-start-pos="1618" data-local-id="dab954b6981b">Scale our current infrastructure and tooling stack to support faster and larger distributed training.</p> </li> <li> <p data-renderer-start-pos="1723" data-local-id="6da6355c88df">Design data strategy to support large scale datasets and efficient storage, ensuring GPUs stay warm.</p> </li> </ul> <h2 id="You-should-apply-if-you:" data-local-id="8dd7eafb-5352-4928-bdb1-3bc8fc536513" data-renderer-start-pos="1828"><strong data-renderer-mark="true">You should apply if you:</strong></h2> <ul data-local-id="a72611fc-69cc-4d66-b54e-45778c988cd3" data-indent-level="1"> <li> <p data-renderer-start-pos="1856" data-local-id="603f9a0c3d6c">Hold an MSc or PhD in Computer Science or STEM field, with a focus on Machine Learning and Deep Learning.</p> </li> <li> <p data-renderer-start-pos="1965" data-local-id="6ba94194216e">Have strong software engineering skills in Python and fluency with modern DL frameworks (PyTorch/JAX/TensorFlow). You don’t just import libraries; you are comfortable writing custom layers, loss functions, and distributed training loops.</p> </li> <li> <p data-renderer-start-pos="2206" data-local-id="c4df54f1606d">Are a clear communicator who can build from complex theoretical concepts and contribute to the company's internal engineering culture.</p> </li> <li> <p data-renderer-start-pos="2344" data-local-id="23cd4273d332">Have a "first-principles" mindset: you enjoy reading the latest AI optimisation blog posts and integrating them into codebases rapidly.</p> </li> <li> <p data-renderer-start-pos="2483" data-local-id="b3a51025a1b7">You have debugged production ML pipelines and can tell a good war story about finding a subtle numerical or performance issue.</p> </li> </ul> <p data-renderer-start-pos="2613" data-local-id="38c4826c4c75"><em data-renderer-mark="true">Note: We operate at an intersection where women, as well as other minority groups, are systematically</em> <em data-renderer-mark="true">under-represented. We encourage you to apply even if you don’t meet all the listed qualifications; ability and impact cannot be summarised in a few bullet points.</em></p> <h2 id="Nice-to-have" data-local-id="dff8be52-1df7-41ab-b00e-65d3bcb2c6d8" data-renderer-start-pos="2879"><strong data-renderer-mark="true">Nice to have</strong></h2> <ul class="ak-ul" data-local-id="2f1cf176-fdf2-4610-ac66-71b3f7b7a33d" data-indent-level="1"> <li> <p data-renderer-start-pos="2895" data-local-id="3207f590b316">You have hands-on experience training models on large-scale GPU clusters, implementing advanced parallelism strategies, and understanding the underlying cross-node communication patterns (NCCL, MPI).</p> </li> <li> <p data-renderer-start-pos="3098" data-local-id="ac9328b80fb1">You have worked with large-scale datasets made of different modalities. Understanding the tradeoffs in locality, encoding, formats, and streaming strategies.</p> </li> <li> <p data-renderer-start-pos="3259" data-local-id="53cf0a79cbc6">You are proficient with workload orchestrators like Slurm, Kubernetes, or Ray, having used them to schedule and manage complex, concurrent training jobs at scale.</p> </li> <li> <p data-renderer-start-pos="3425" data-local-id="3937b80ec1bc">You understand GPU architecture at a low level: memory hierarchies, warp execution, and what makes a GPU suited for training versus inference workloads.</p> </li> </ul> <h2 id="Join-Helsing-and-work-with-world-leading-experts-in-their-fields" data-local-id="569e4b6d-0818-4563-87b8-f93c7baeb9aa" data-renderer-start-pos="3581"><strong>Join Helsing and work with world-leading experts in their fields</strong></h2> <ul data-local-id="13f2ff97-2878-4d88-ae