**Problem Solving (Integrations) - Technologies**
**1.
** There is a myriad of small problems to be tested and solved.
These are predominately involved where we have unknowns involving interfaces or integrations between systems.
**i**.
e.g.
Running MATLAB code in Databricks, loading data from Snowflake into databricks, building an API that can read data from databricks.
**ii**.
Yes, most of this currently revolves around Databricks, but a databricks expert isn't necessary.
Even databricks experience isn't necessary.
**iii.
** I believe the conceptual architecture and component choices are solid and built on a strong foundation, we just need to start wiring the components together.
**2.
** Most integration challenges like this can be solved by finding, reading, and following documentation and then troubleshooting any issues that arise.
**2.
** Prototyping (Fail Fast)
1.
Be willing to fail fast and ask for help.
We won't learn what will and won't work without being willing to try something might not work.
2.
Take the Edison approach - I haven't failed, I've just found 10,000 ways that won't work.
3.
Start with small examples and work your way up from there.
Worry about solving problems incrementally, we don't need to solve every problem on day 1.
**i.
** 0 -> 1 is hard, 1 -> n is easy
**ii.
** Start with a table that has 2 columns and add more columns later.
**iii.
** Start with "Hello world" examples when testing new integrations, then add functionality later.
**iv.
** Start with a simple model y = mx + b and add more complexity later
**3.
Communication Skills**
1.
There is a lot of uncertainty in this space and strong communication skills will be necessary to find answers to problems.
**i.
** Working with key business partners to learn and understand the domain knowledge.
**ii.
** Working with architects to understand how and why we're planning to build the solution.
**iii.
** Working with project managers to understand timeline expectations and build a plan of attack.
4.
** Python**
1.
Strong python skills will be necessary here as it is the common language between EMTEC and EMIT.
**2.
** Both scripting and API experience will be useful.
We'll need to create job code that can run reliably in Databricks and we'll need to build APIs that can expose the results of models.