You are spot on that sometimes we just overcomplicate models and sometimes it’s best to go with something explainable and deterministic but less accurate as opposed to more accuracy but complicated.
935 karma · joined March 7, 2012
You are spot on that sometimes we just overcomplicate models and sometimes it’s best to go with something explainable and deterministic but less accurate as opposed to more accuracy but complicated.
The environment in the US is just at a different level compared to everywhere else.
I’ve worked in the U.K. and speaking with many data scientists from other parts of Europe, they say that the real data science work is just non-existent elsewhere and pays less.
Avora (https://avora.com/product/) and Thoughtspot (https://Thoughtspot.com) all have the root cause capability
Additionally, what type of support does Huawei/Alibaba offer for startups?
Still to be determined with the client but yes, the sensors will be installed in a factory. Client already has some sensors to detect things but am thinking we may need other sensors in case the existing ones do not give enough information. Additional sensors could be temperature/humidity, wear indicators, vibration and sound.
Are there any websites I can visit to view these sensors made by GE, Siemens, etc.? Could sensors from element14 and adafruit be used in an industrial setting?
If you're interviewing with larger companies, they generally tend to have more resources to hire those other skillsets so your focus as a data scientist is to create models.
Aside: can I ask for some advice on software requirements? We’re designing our solution to meet IT requirements and want to get under the skin of CIO concerns.
Tools: Avora, FiveTran, Informatica, Excel, Alteryx, Datameer, Tableau Prep, etc. ... list could go on and on.
You know your target customers but perhaps you also need to understand how they buy software. For us, a website and the usual inbound tactics are great but our target customers will expect sales reps, account execs, etc.