Agentic AI Engineer (Engineering, Web&Mobile Team)

We’re always looking for the ones truly passionate about their work. If you are amongst them, you can rest assured there is a place for you in eMAG. We’ve grown very fast and are determined to keep doing so. What brought us here is our desire for continuous evolution and practical results. More than 6000 colleagues are part of eMAG Teams . We strongly believe in people's development and therefore every year we invest more and more energy and resources to remain an organization that is constantly learning. We want to ensure that you’ll have the most talented colleagues and the proper environment to grow and achieve great results, to become what you desire on a personal and professional level. Join us, grow faster ! Agentic AI Engineer (Engineering , Web&Mobile T eam) As an Agentic AI Engineer , you will build AI-powered workflows that accelerate engineering teams and improve developer productivity. You will orchestrate LLMs and AI agents into reliable, production-grade tooling that operates at scale. The role requires strong software engineering/ architecture expertise, combined with a product mindset and the ability to adapt quickly in the rapidly evolving agentic AI landscape. What you’ll have to do: Design and ship agentic workflows across the dev lifecycle — auto-documentation, test generation on MRs, parity testing for migrations, review augmentation; Build reusable platform components: MCP servers, tool integrations, prompt/ orchestration libraries other teams can self-serve; Own eval & observability for AI flows (Langfuse, LLM-as-judge, regression suites) so outputs are trustworthy, not just impressive demos; Deploy flows as production services (Docker, K8s) with proper CI/CD and cost/ latency awareness; Drive adoption at scale — measure impact via developer-productivity metrics (e.g. getDX), iterate on what actually moves throughput and cycle time; Keep pace with and adapt to the fast-moving agentic landscape — continuously evaluate new tools, models, and patterns, fold the useful ones into how the org works, and bring pragmatic choices. What makes you a good fit: Strong programming fundamentals and a track record shipping production-grade software (Python is the primary language of the agentic ecosystem); Hands-on with agentic coding tools (Claude Code, Cursor, Aider, Copilot) and having built at least one real agentic flow in production; Working knowledge of LLM tool calling and MCP (Model Context Protocol); Experience with one or more orchestration frameworks (LangGraph, LangChain, or equivalent custom orchestration on LLM provider APIs); Comfortable with cloud, Docker, and Kubernetes; Solid Bash / Shell scripting. What makes you stand out: Multi-agent / agent orchestration patterns at scale; RAG, LLM eval/ observability, Langfuse, TypeScript / JavaScript, Go, PHP; Direct experience working across multiple LLM providers and models (Anthropic, OpenAI, DeepSeek, open-weight models, and others) and comfortable switching as the landscape evolves. What we’ve prepared for you: Medical subscription: Medicover or Regina Maria. A flexible budget that you can invest in yourself as you wish: meal tickets, holiday tickets, cultural vouchers, private pension, foreign language classes, eMAG, Fashion Days, Therme & Genius, membership to different gyms or even professional development classes. Different discounts from our partners: banking, mobile, dental medicine or wellness. Access to the Bookster library and free credits on the Hilio psycho-emotional health platform. An accelerated learning environment, with access to over 100.000 curated online resources and platforms, learning academies and development programs. New headquarters, where sleek design, natural light, and versatile spaces create an energizing and comfortable environment for hybrid work. #LI-remote #midsenior Curious to find out more about the next step in your career? Apply now and if your experience is relevant for the role you wish, we will give you a call for more details! Also, here you can find our confidentiality policy if you want to consult it.