This paper presents an intent-driven orchestration approach for serverless applications deployed across the computing continuum. The approach aims to optimize the scheduling and autoscaling of serverless functions to meet Service Level Objectives (SLOs) while minimizing resource usage, communication overhead, and energy consumption. The proposed solution uses a Reinforcement Learning (RL)-driven autoscaling mechanism that incorporates a reward function to optimize resource utilization and ensure timely autoscaling. Additionally, a scheduling optimizer is introduced to determine the optimal placement of function chains across the continuum, minimizing link utilization, transformation cost, and power consumption. The solution is evaluated against state-of-the-art techniques in a small-scale testbed, demonstrating a 1.52% QoS violation with a slight increase in deployed resources and significant reductions in cost and power consumption. The system architecture includes a workload estimator, RL-based autoscaling, and a scheduling optimizer, all working together to achieve the deployment intent. The approach is validated through a multi-objective optimization problem that considers the system's performance, energy efficiency, and cost. The results show that the proposed method outperforms existing techniques in terms of QoS, cost, and power consumption. The paper also discusses related work, including intent-driven orchestration mechanisms, autoscaling in serverless computing, and distributed function scheduling in the computing continuum. The study highlights the importance of intent-based orchestration in managing serverless applications across the computing continuum, emphasizing the need for efficient resource scheduling, autoscaling, and energy-efficient placement of functions.This paper presents an intent-driven orchestration approach for serverless applications deployed across the computing continuum. The approach aims to optimize the scheduling and autoscaling of serverless functions to meet Service Level Objectives (SLOs) while minimizing resource usage, communication overhead, and energy consumption. The proposed solution uses a Reinforcement Learning (RL)-driven autoscaling mechanism that incorporates a reward function to optimize resource utilization and ensure timely autoscaling. Additionally, a scheduling optimizer is introduced to determine the optimal placement of function chains across the continuum, minimizing link utilization, transformation cost, and power consumption. The solution is evaluated against state-of-the-art techniques in a small-scale testbed, demonstrating a 1.52% QoS violation with a slight increase in deployed resources and significant reductions in cost and power consumption. The system architecture includes a workload estimator, RL-based autoscaling, and a scheduling optimizer, all working together to achieve the deployment intent. The approach is validated through a multi-objective optimization problem that considers the system's performance, energy efficiency, and cost. The results show that the proposed method outperforms existing techniques in terms of QoS, cost, and power consumption. The paper also discusses related work, including intent-driven orchestration mechanisms, autoscaling in serverless computing, and distributed function scheduling in the computing continuum. The study highlights the importance of intent-based orchestration in managing serverless applications across the computing continuum, emphasizing the need for efficient resource scheduling, autoscaling, and energy-efficient placement of functions.