We're looking for a Founding AI Engineer to help build Akaion's core intelligence layer from day one. You'll design, ship, and scale production-grade AI systems that power our next-generation products. Join Akaion and shape the future of our AI-first platform. You'll work at the intersection of product, research, and engineering to turn cutting-edge AI into real-world impact.
What You'll Do
Design and build end-to-end AI systems (from data ingestion to production inference) Develop and deploy LLM-powered agents and multi-agent architectures Build scalable Gen AI services and APIs for product integration Design prompt engineering, tool calling, function routing, and orchestration layers Optimize model performance, latency, and cost Deploy AI workloads on Google Cloud Platform (GCP) (Cloud Run, GKE, Vertex AI, storage, networking) Build monitoring, logging,
evaluation and feedback loops for AI systems Collaborate directly with product and founding team on roadmap and architecture decisions What We're Looking For Strong experience with Python for AI/ML and backend systems Hands‐on experience with LLMs, Gen AI frameworks (Lang Chain, Llama Index, Open AI SDKs or similar) Experience building RAG systems and vector search pipelines Knowledge of agent-based architectures and AI orchestration patterns Experience with cloud infrastructure (GCP preferred) Familiarity with Docker, containerized AI workloads, and scalable deployments Solid understanding of REST APIs, async systems, and production engineering Ability to move fast and own systems end-to-end Nice to Have (Big Plus) Experience with Vertex AI, Hugging Face, fine‐tuning or model serving Knowledge of streaming pipelines, data engineering, or MLOps Experience with evaluation frameworks for LLMs Background in startup or early‐stage environments Interest in AI agents, autonomous systems, and AI‐native products Why Akaion?
You'll be a founding technical pillar Massive ownership and architectural influence Real product imp
📌 Founding ai engineer (Milano)
🏢 Akaion
📍 Milano