Applied AI Engineer

You will work with every part of the company: leadership, business development, operations, and engineering teams across hardware, software, systems and test. You will learn how each team actually works, find where AI creates real leverage, prove it with fast experiments, and turn the winners into tools people use every day. You are a builder who understands both code and business processes. You are an early adopter: you were using AI terminal agents and MCP when they were new, and you read changelogs instead of summaries. You also know that in a defense company, the right answer is sometimes a frontier model in the cloud and sometimes a model running on our own hardware with no internet at all. You will be the person who knows which, and builds both. This is a hands-on role with wide scope across the company, with room to grow as the AI function scales. Skapion is building a Native Counter-Swarm Platform, and AI is part of how the whole company designs, builds and operates. This role owns that effort, from the engineering lab to the back office. • Real impact: Your systems will be used every day, by everyone from engineers in the lab to the leadership team. • Hard problems: Apply AI to real hardware, real data and real constraints, not just chat interfaces. • Freedom to shape: Help define how a young defense-tech company works, with real influence on how AI is adopted across teams. • Room to grow: Start hands-on, with the opportunity to build and lead an AI team as the company scales. • Deliver the roadmap and find what's next: Implement the AI initiatives prioritized by leadership, and proactively spot new opportunities across engineering, R&D, business development and operations. Turn the best ideas into clear, simple proposals for management, with expected impact and cost. • Understand the business first: Sit with process owners, from procurement and finance to test engineers and program managers. Map how work is done today, find the bottlenecks, and design AI workflows that fit how people really work. • Own it end to end: Take each initiative from concept through design, build, pilot, validation and handover. Every capability ships with documentation, training and a named owner, and you keep improving it after launch. • Prove it fast, measure it honestly: Run proofs of concept with LLMs, agents and automation. Measure results against the current process together with team leads, and decide what graduates to production and what gets dropped. • Build the AI platform: Develop the shared foundation teams build on: agentic workflows, MCP servers, LLM integrations, RAG pipelines, and connections to company systems such as M365 and business applications. • Deploy in the right environment: Choose where each workload runs based on data sensitivity, cost and capability, from managed cloud APIs to self-hosted models on private infrastructure and disconnected networks. Build systems that can move between them. • Bring AI into the lab: Work with engineering teams to automate test benches and development loops, including instrument control, build-and-flash pipelines and automated analysis of test data, with engineers always approving the result. • Keep it safe and observable: Build evaluation, monitoring, cost controls and guardrails into every system. Set data-handling and model-access practices suited to a defense environment. • Drive adoption: Train teams, support internal AI champions, maintain a shared library of skills and prompts, and report usage and impact to leadership. • Stay ahead of the curve: Track new models, frameworks and research. Test what matters, ignore the hype, and share what you learn. • 5+ years of professional software engineering, with a strong backend focus and products shipped to production. • A proven record of architecting and delivering AI-powered features or platforms that people use daily, not just prototypes. • Experience analyzing business or engineering processes with their owners, and turning them into automated workflows with measurable impact. • Hands-on production experience with leading commercial LLM APIs, including prompt engineering, tool use and agentic workflows. • Hands-on experience running open-weight models on private infrastructure, including inference serving, quantization and hardware sizing. • Practical experience building RAG systems: chunking, retrieval, re-ranking and evaluation. • Experience with MCP, either building servers or connecting agents to internal tools and data. • Strong Python, working proficiency in TypeScript/JavaScript, and familiarity with the AI/ML ecosystem (e.g., PyTorch, Hugging Face). • Experience deploying in both cloud and on-premise environments: containers and Kubernetes, CI/CD, and self-managed data stores including vector databases. • Daily use of AI coding agents in real engineering work, with the judgment to review and correct their output. • Excellent communication skills: you can explain complex AI technology to management simply and clearly, in writing and in presentations, and earn the trust of non-technical colleagues. • Self-direction: you can scope your own work, make the call and deliver without a playbook. • Eligibility to obtain and maintain a security clearance. • Fluent Hebrew and English. • Deploying AI in air-gapped or security-sensitive environments, including model vetting and offline updates. • Edge AI: optimizing and deploying models on embedded GPUs (e.g., NVIDIA Jetson) with ONNX or TensorRT, including validation and regression testing. • Lab and test automation: instrument control (SCPI/PyVISA), hardware-in-the-loop setups, or working alongside embedded firmware teams. • Familiarity with systems engineering practice: requirements and traceability, FMEA, configuration and change management. • Background in computer vision, sensor fusion, control, autonomy or other applied research. • AI observability and evaluation frameworks, and monitoring models in production. • A strong understanding of LLM internals or fine-tuning. • Integrating AI with enterprise systems such as M365, SharePoint, ERP or CAD tools. • Experience in defense, aerospace or other regulated industries. • Published research, open-source contributions, or side projects that show original AI work.