Embodied AI Engineer at Mecka.ai

THE ROLE We're hiring an AI Robotics Engineer based in New York City to help bring robotic intelligence from models into real-world execution. This is a highly applied role focused on deployment, integration, and system performance, while also working deeply with learning-based policies (RL), simulation systems, and control frameworks. You'll work on connecting perception, world models, policy learning, and control into systems that actually function reliably on real robots. This role is ideal for engineers coming from top robotics companies (e.g. humanoids, autonomy, manipulation) or top academic labs, who want to work on real-world systems and move quickly from idea → deployment. RESPONSIBILITIES MODEL INTEGRATION & SYSTEM EXECUTION - Integrate vision, multimodal, world models, and policy networks (RL / imitation learning) into robotic systems - Translate model outputs into real-world control signals and actions - Build the glue between perception → prediction → decision → actuation PERCEPTION & STATE ESTIMATION - Develop and improve perception pipelines - Work on pose estimation, scene understanding, and object interaction - Build accurate state representations for downstream control and policy learning - Ensure robustness in real-world, noisy environments CONTROL, MANIPULATION & MOVEMENT - Implement and tune control systems for manipulation and locomotion - Apply classical control methods (PID, MPC, trajectory optimization, state estimation) alongside learned policies - Work on stability, smooth motion, and real-time responsiveness - Handle real-world constraints like latency, drift, and noise REINFORCEMENT LEARNING & POLICY DEVELOPMENT - Work with reinforcement learning (RL) and imitation learning to train robot policies - Design reward functions, training setups, and evaluation frameworks - Leverage offline + online data to improve policy performance - Collaborate on integrating learned policies into real-world systems SIMULATION & SIM-TO-REAL - Use simulation environments (e.g. Isaac, MuJoCo) to train and validate policies - Work on sim-to-real transfer, domain randomization, and system identification - Use simulation to model environments, dynamics, and edge cases - Leverage real-world data to improve generalization and reduce sim gaps DEBUGGING & DEPLOYMENT - Debug full-stack issues across models, learning systems, software, and hardware - Work directly with robotic systems in real-world environments - Diagnose failures across perception, policy, and control layers - Iterate quickly to improve system performance and reliability WHO YOU ARE REQUIRED EXPERIENCE - 4+ years of experience in robotics, AI/ML, or related fields - Strong programming skills in Python (C++ strongly preferred) - Experience working on real robotic systems or deployments - Solid understanding of robotics fundamentals, control systems, and/or RL - Experience integrating models or policies into production or real-world systems TARGET BACKGROUNDS - Experience at leading robotics companies (e.g. humanoids, autonomy, manipulation systems) - Experience in applied robotics labs or systems-focused research environments STRONG SIGNALS - Experience with reinforcement learning (RL), imitation learning, or policy optimization - Experience with world models, dynamics modeling, or predictive systems - Strong background in classical controls (PID, MPC, state-space methods) - Experience with dexterous manipulation, hands, or locomotion - Experience with simulation environments (Isaac, MuJoCo, etc.) - Familiarity with ROS or similar middleware - Experience with vision-language-action (VLA) or multimodal models - Experience working with large-scale real-world datasets YOU ARE - Highly practical and execution-focused - Comfortable working across learning-based and classical robotics systems - Strong at debugging complex, real-world systems - Able to move quickly from prototype → deployed system - Motivated by building robots that actually work outside the lab WHY THIS ROLE - Work directly on real-world robotics systems with leading AI labs - Operate at the intersection of data, learning systems, and physical robots - High ownership and direct impact on deployments and performance - Solve hard problems in manipulation, control, and embodied intelligence - Help shape how robotics systems are trained and deployed at scale COMPENSATION - Base Salary: $160,000 – $220,000 USD - Equity offered in addition to base salary