Taylor Murphy
Taylor Murphy
@tayloramurphy.dad
Arch Data (@arch.dev) Founder/CEO Meltano maintainer 🐉 ex-Data @ GitLab, Concert Genetics PhD @ Vanderbilt Paid Data Influence / Actor aka "The Peter Drucker of Data Shitposts" Husband and Dad 👦👦👧
3.8K followers295 following593 posts
Taylor Murphy reposted
This post from Ramp is really cool. Tickles all my "data orderliness" spots engineering.ramp.com/industry_cla...
I love this article world.hey.com/jason/we-inc...
I enjoyed this post from my good friend Emilie! open.substack.com/pub/emilie/p...
How did Elastic adding a more open source license not make the rounds when they announced it?? www.elastic.co/blog/elastic...
see? @jlowin.dev was right - you should've changed your title to analytics engineer!
so I use SwiftKey on ios and it's better than the native keyboard but it still sucks. the swipe gesture doesn't spell very common words. "I'm at the aurora" instead of "airport" 🙄
Is it just me or is the new iOS keyboard horrible? I've made so many more typos lately, and the autocorrect is wildly off way more often than usual!
Taylor Murphy reposted
I got this working and it feels so magical. Was able to update some data pipelines just by asking Claude. FastMCP was great - the browser based debugger helped tremendously. I had Claude write the whole server . py file (with a few changes) after feeding it our openapi spec and the fastMCP readme.
I got Claude loaded up with some tools via MCP. Very cool experience. I want to test FastMCP now (hat tip @jlowin.dev) to see if I can connect it to the @arch.dev api and do some fun things 👀
I got Claude loaded up with some tools via MCP. Very cool experience. I want to test FastMCP now (hat tip @jlowin.dev) to see if I can connect it to the @arch.dev api and do some fun things 👀
I think nearly all tech/data advice should be couched in the size of the org. When I reflect on the "Run your data team like a product team" advice, we should've been way more explicit about org size (i.e. more relevant for >1000)
Working at a company that's gotten pretty sizable headcount-wise (~8000ish), I'm finding a lot of advice for data teams is geared towards smaller companies. Things that worked for me when I was at small orgs just don't most of the time. Build/buy decisions, time needed to change infra, etc.
this is an interesting thread to read. clear overexuberance by execs and maybe some over promising by a dbt sales eng (I'm guessing) and it ends in disaster. www.reddit.com/r/dataengine...
yeah so I'm never letting my kids play Roblox hindenburgresearch.com/roblox/