I build end-to-end Computer Vision and ML systems, from model selection and training pipelines through to deployment on CPU and edge hardware. My background spans applied research and hands-on experience with sensors and prototypes across four RWTH institutes, so I can navigate both a paper and a production codebase.
What I bring to a team:
-
Turning research viable : 4× runtime speedup via parallelisation, automated mesh generation pipelines, and a prototype for automating Textile QA coupled with a Neural Network
-
Automating annotation & cutting costs : built SAM-based semi-automated labelling tool and a full UDA pipeline (12 models, 150+ papers reviewed) to reduce manual labelling in deep learning-based defect detection at Fraunhofer IPT
-
Inference under constraints : deployed YOLOv10 at 46 FPS on CPU-only hardware via ONNX Runtime, building a zero-copy DeepStream pipeline targeting NVIDIA Jetson
Master of Science in Robotic Systems Engineering RWTH Aachen University, Germany | Grade: 2.1/4.0 (German scale : equivalent to Good)
Bachelor of Technology in Production Engineering (with Honours) College of Engineering Pune, India | GPA: 8.33/10.0
A short deck walks through projects done at various RWTH Institutes while working as a student research assistant (Studentische Hilfskraft).
Previous Works (click to open the PDF) |
Human-in-the-loop annotation workflow |
- Programming Languages: Python, C++, C, MATLAB, Julia
- ML Frameworks: PyTorch, TensorFlow, scikit-learn
- Libraries: NumPy, Pandas, OpenCV, Matplotlib
- Robotics Middleware & Backend: ROS2, FastAPI, SQL
- MLOps & Deployment: MLFlow, Docker, ONNX
- Version Control & CI/CD: Git, GitHub Actions
- Simulation & 3D Modelling: VeroSim, Gazebo, Blender, SolidWorks
- 📧 Email : ambuj.choudha@rwth-aachen.de
- 📍 Based in Aachen, Germany
Open to full-time roles in Computer Vision, ML, and Robotics Software integration
Interested in collaborations on Computer Vision, Robotics, or ML projects? Feel free to reach out!