"...anomaly detection solutions ... leveraging an ensemble of Deep learning and LSTM models. Natural Language Processing for entity, topic clusters and relationship extraction. Text Analytics in human generated tickets and correlation with event tickets for event noise reduction. Apply Natural Language Classification and RNN algorithms to automatically route tickets... Text mining, message clustering / templatization, Logs to metrics, anomaly detection, event annotation and sequencing... for each mainframe batch job ... Identify Anomalies ... using sequence mining techniques"
In all it is an interesting product but it is going to take a lot of people with different skills to make that happen. Many of those areas such as "entity and relationship extraction" and "anomaly detection solutions ... [for] event noise reduction" are still almost pre-paradigmatic from the viewpoint of a working engineer.
The best interpretation I have is that they are hiring a large platoon or small company of software developers, data scientists, project managers, you name it. They are probably using a distributed version of UIMA that spins up at least one container per dev in the production system (because "microservices".)
The person they are hiring here is in charge of keeping that monster going at the K8s level. They may need to settle for hiring 12 people with 1 year of experience, but that would blow their budget -- if the whole team worked that way it would get bloated to a mid-sized or large battalion.
I don't see why they don't just run the system on one of the mainframes it is monitoring. With Parallel Sysplex, Workload Manager, etc. IBM had better stuff than VMWare, Docker, K8S, Zookeeper in the 1990s when they made the transition from bipolar to CMOS and had to go parallel to make up for the single-thread performance loss. z15 mainframes are just crammed with PCIe slots so they should have no problem attaching a tensor accelerator to one.
Trouble is, people who know how to administer mainframes are even harder to find than K8S experts.