TechVill is an IT company and a partner of VkusVill for digital solutions development.**

We are responsible for the development of mobile and web applications, business process automation, artificial intelligence, devops, and information security for VkusVill.

Our solutions are used by over 1,000,000 clients and employees of VkusVill.

Currently, we are looking for a Robotics Software Engineer to join the Darkstore Robotization team. The role will involve participating in the robotization and automation of processes at VkusVill.

Responsibilities:

• Implementation and adaptation of manipulation and navigation algorithms: deployment, configuration, and refinement of ready-made solutions for trajectory planning (Motion Planning), obstacle avoidance, object grasping (Grasp Planning), and manipulator control in conjunction with computer vision systems;
• Integration of computer vision software: embedding and optimizing pre-trained models and classical CV algorithms to solve applied tasks: object detection, segmentation, classification, and 6D pose estimation. The goal is to take a working model prototype and integrate it into the robotics stack with guaranteed performance;
• ML model support (ML-Ops on board the robot): packaging models for production (ONNX/TensorRT), ensuring efficient inference on onboard compute units, and monitoring the correct operation of models on the robot;
• Development of middleware on ROS 2: creating and configuring nodes, topics, services, and actions as the robot's unified nervous system, connecting control, perception, and actuator modules;
• High-performance data processing: optimizing speed-critical modules (e.g., filtering and segmentation of point clouds from lidars/Depth cameras) using C++ to achieve real-time performance;
• Working with data: collecting, labeling, and preprocessing data from sensors (RGB-D cameras, lidars, tactile sensors) for retraining and fine-tuning integrated models;
• Full integration cycle: from launching in a simulator and debugging a prototype to deployment, field testing, and supporting code on real robots.

Our Technology Stack:

• Languages: C++ (17/20), Python 3.x;
• Robotics Frameworks: ROS 2 (Humble), MoveIt 2, CuRobo, Navigation 2 (Nav2), Autoware;
• Computer Vision and 3D: OpenCV, PCL;
• ML and Optimization: PyTorch, ONNX Runtime, TensorRT;
• Simulation: Isaac Sim, Gazebo (Ignition), Mujoco, Genesis;
• Infrastructure and Tools: Docker, Linux, Git, CI/CD (GitLab CI/Jenkins);
• Sensors: RGB-D and stereo cameras, lidars, tactile sensors.

Requirements:

• 2+ years of commercial experience in C++ development, including code optimization and multithreading;
• Practical experience integrating systems in ROS 2 (creating custom nodes, working with actions, launch files);
• Confident solving applied Computer Vision problems using OpenCV, understanding how to choose a detector/segmenter for a task and "attach" it to a robot;
• Experience working with 3D data: point clouds (PCL) or RGB-D images;
• Understanding and practical experience applying robotics fundamentals: forward/inverse kinematics, working with coordinate systems (TF2);
• Proficient work in Linux OS (advanced user level) and Git proficiency;
• Practical experience packaging and running ML models (understanding what ONNX/TensorRT are, ability to measure inference speed and write an efficient data preprocessing pipeline);
• Ability to write clean, modular, well-documented code.

Desired:

• Experience with MoveIt 2, CuRobo (configuring kinematics, collisions, integration with real robots and simulators);
• Experience with Navigation 2 (Nav2), Autoware;
• Experience launching and debugging robots in Isaac Sim or Gazebo simulators;
• Understanding of neural network architectures (YOLO, PointNet, EfficientNet) and Transfer Learning principles to a sufficient level for fine-tuning models for your specific task.

Conditions:

• Work in an accredited IT company.
• Hybrid; Work on-site is required during key project stages. Readiness for periodic visits is expected.
• Official employment from the first day of work and curator support during adaptation.
• Transparent development system: clear grades, internal and external training, individual development plans, and competency matrices.
• Friendly culture and supportive management.
• Compensation for medical services, mental well-being, sports, team-building events, and AI assistant usage.
• 15% discount on purchases at VkusVill.
• Social responsibility: we encourage blood donation, provide financial assistance upon childbirth.
• Partner program "Green Light": you can receive up to 50,000 rubles for referring specialists.