Our company is actively **hiring** for a Simulation and 3D Reconstruction Algorithm Engineer through our campus **recruitment** program, offering outstanding **employment opportunities** in robotics and AI technology. This position provides the chance to work on cutting-edge SLAM algorithms and neural rendering techniques while developing humanoid robot perception systems. Successful candidates will join our Shenzhen team to build simulation environments and optimize algorithms for embedded platforms. If you're beginning your professional **careers** in robotics perception, this represents an excellent opportunity in the emerging humanoid robot industry.

The ideal candidate should meet the following requirements:

1. Familiar with the classic SLAM framework (ORB-SLAM3, VINS-Fusion, LIO-SAM), understand the front-end (feature matching, optical flow), and back-end optimization (BA, posture map optimization).
2. Master 3D point cloud processing (PCL/Open3D), multi-view geometry (MVG), and depth estimation (MonoDepth).
3. Familiar with CUDA acceleration (such as GPU-based TSDF reconstruction) or embedded optimization (ARM NEON/TensorRT).
4. At least one mainstream simulation platform (Isaac Sim, Gazebo, CARLA, Unity3D) has been used, and it can build a basic simulation environment.
5. Familiar with ROS/ROS 2, able to develop sensor data interfaces (such as Camera/IMU/LiDAR simulation).
6. Proficient in C++/Python, master the characteristics of modern C++ (11/14), and be able to write high-performance algorithm code.
7. Familiar with PyTorch/TensorFlow, and experience in deep learning model training (such as CNN, Transformer).

Familiar with the following research directions (at least one)：
1. Application of NeRF/3DGS (3D Gaussian Splatting) in SLAM or reconstruction.
2. Dynamic SLAM (such as DynaSLAM, FlowFusion).
3. Semantic SLAM (such as MaskFusion, CubeSLAM).
4. Sim2Real (simulation to real migration learning).
5. AIGC related experience (priority is given)：
6. Use Diffusion Models/Stable Diffusion to generate synthetic training data (such as texture enhancement, scene synthesis).
7. Robot scene understanding (such as natural language instruction navigation) based on LLM (such as GPT-4, LLaMA) or VLM (such as CLIP, BLIP-2).
8. Familiar with generative 3D modeling (such as DreamFusion, Magic3D, Shap-E).

Check out the full job listings here