As I know integration with Hadoop, Spark, Kafka, ElasticSearch and any tool which uses JVM is not a piece of cake. The same situation with Python (especially with projects and tools, which uses 2.7 version) to port someone's code to IronPython and back (if I want to share my changes with other people) is not a piece of cake as well. I'm not talking about Scala, Haskell, and Erlang. Sometimes the only option to integrate with them - design endpoint(-s) with declared API (which could be also a head pain as well). If you were me how would you try to resolve integration issues for each case?

Scott, hello from Ukraine! In my opinion, stack in Java is highly ready for set up BigData high load architecture (I mean Hadoop, Spark, Kafka, ElasticSearch and so on). DataScience's stack is mostly associated with Python and R now. When some projects need functional development - most of all they are looking for Scala, Haskell, Erlang and even Python engineers (but not F#). Could you explain, where is .NET doing better (except syntax)? How to compete with that tools?

What really helped me - I started attending boxing training. At first, it was nothing, but fitness with(and without) gloves, a lot of running, a lot of jumping and so on. But then body got a pretty good shape and my long sitting at office became not so painful. I realized - it was what I need. I has been attended boxing training session twice a week for a last three years. And my back's muscle corset became so strong that I can sit a couple of hours without break and don't feel tired because of that. As a bonus I can knock out anyone who is standing on my way. ;-) P.s.: please, do not forget to stretch your body after each training session - it's very important (to give your muscles some sort of relax command)!