Python: Machine Learning models OOM issue in Kubernetes deployment
Budget: €1,500 – €3,000 EUR
I am looking for a Python expert to help resolve an out-of-memory (OOM) issue with machine learning models in a Kubernetes deployment. The specific models being used are Isolation Forest, XGBoost, and Neural Networks.
Requirements:
- Strong experience in Python and machine learning
- Experience using multi-processing in Python
- Expertise in Kubernetes deployment
- Familiarity with memory and CPU limitations in Kubernetes
The expected outcome is to resolve the OOM issue and ensure smooth functioning of the machine learning models in the Kubernetes environment.
We are not sure whether the issue stems from an inherent memory leak in our application, or whether our Kubernetes deployment is not setup correctly. What we are seeing is: Our python processes consume more memory then what they have available in Kubernetes. The memory seems to go gradually upwards, and the garbage collector does not seem to kick in.
Requirements:
- Strong experience in Python and machine learning
- Experience using multi-processing in Python
- Expertise in Kubernetes deployment
- Familiarity with memory and CPU limitations in Kubernetes
The expected outcome is to resolve the OOM issue and ensure smooth functioning of the machine learning models in the Kubernetes environment.
We are not sure whether the issue stems from an inherent memory leak in our application, or whether our Kubernetes deployment is not setup correctly. What we are seeing is: Our python processes consume more memory then what they have available in Kubernetes. The memory seems to go gradually upwards, and the garbage collector does not seem to kick in.