1. Deploy Machine Learning/Deep Learning models to Production
-
Convert models developed by Data Scientists into complete products through API/microservice, ready for integration with business systems.
-
Build and operate realtime or batch inference systems suitable for each use case.
-
Ensure models operate stably in Production environment, meeting requirements for SLA, latency and scalability.
2. Manage model lifecycle (MLOps)
-
Design and operate CI/CD pipeline for ML, from training, testing to deployment and retraining.
-
Automate model lifecycle: training – testing – deployment – monitoring – retraining.
-
Manage model registry, versioning, rollback and change control according to MLOps standards.
3. Monitor and maintain model performance
-
Track model performance after deployment; detect data drift, concept drift and signs of quality degradation.
-
Build monitoring and alerting systems when models fail to meet requirements.
-
Collaborate with Data Scientists to improve models based on real-world data and system feedback.
4. Optimize performance and AI infrastructure
-
Optimize training and inference speed, ensuring performance in Production environment.
-
Optimize CPU/GPU and system resource usage, contributing to operational efficiency.
-
Participate in evaluating and proposing AI architecture and infrastructure suitable for the Bank's direction.
5. Technical and business collaboration
-
Collaborate with Data Engineer, DevOps and IT in deploying, integrating and operating AI solutions.
-
Collaborate with business units in implementing AI models into systems.
-
Participate in standardizing AI architecture and ML/MLOps platform across the Bank.
