Marc

post-academic data analyst / software dev. currently at fintech startup. code manages other ppls money.

[vicki @vboykis]

Compliment a data professional in one tweet.

[Senior Oops Engineer @ReinH]

holy, and I cannot stress this enough, shit

[Marc @marc_data]

@patio11 cc
You might appreciate this story.

[George Mack @george__mack]

FBI + DEA was outperformed by one guy *using Google in his spare time*.

The point of this thread isn't to debate the ethics of drugs being illegal.

The point is to highlight the power of an individual with:

1. High agency
2. Internet connection
3. Thinking outside the box

[Marc @marc_data]

@lhochstein This sounds less hypothetical and more a running diary.
If it's me,
Try to take some extra notes, so that when the local experts do become available, they have more info to work with.
But it's been 24 hrs, so you resolved this situation?

[Sean J. Taylor @seanjtaylor]

Some of my biggest regrets stem from being too intimidated to try to learn things I thought would be too difficult. I wasted a lot of time talking myself out of things rather than just giving them a shot.

[Chip Huyen @chipro]

Guys, hear me out. What if instead of showing students that same convnet picture in 10 different classes, some classes teach ML students engineering principles? Idk, something like unit test, CI/CD, performance profiling, typing, encoding, databases, etc.?

[Bert Chan @BertChakovsky]

A nice interactive tutorial of critical threshold in complex systems. In the simplest form it's cellular automata with random rules, but can be used to easily explain real world cases e.g. herd immunity.

https://t.co/P5olpZ6hGD

[Julie Zhuo @joulee]

Researchers, data analysts, VCs can gain product sense through pattern-matching from large sets of data. PMs, engineers, marketers develop it in the trenches of repeated shipping and iteration.

[Auren Hoffman @auren]

number of employees of a well-run 100-person start-up that care about the long-term company: 60

number employees of a 10,000-person company that care about the long-term company: 6

this is why start-ups can win.

[Aaron Richter @rikturr]

Fave quotes/comments from #DataMishapsNight:

"I hate it when my constants turn out to be variable"

"my boss had no idea it was going to be hard"

"please find me some data to fit this story I'd like to tell"

"I am going to be huge”

"spicy extrapolation"

[Arik Friedman @arikf]

Everybody make mistakes. Coming clean about mistakes is always uncomfortable, but integrity is the data analyst's greatest asset.
#DataMishapsNight

[Ibraheim @ibra_0707]

Two things I've learned:

Way too much competition for entry and associate engineer positions

specialty matters

DON'T under value your niche skillset

DO undervalue your general skills (unless your some wiz)

Learn to sell yourself and think on your feet

[Charity Majors @mipsytipsy]

i'll help! ✨start by using CD for any new deploy targets✨

it can be painful to CD-ify a mature, "stable" service, and the pain can be politically unacceptable; but it is utterly painless to grow up with CD from day one (and has a way of turning doubters into vocal converts) ☺️

[Marc @marc_data]

!

[Ben Sigelman @el_bhs]

11/ So, ultimately, of course there were ∞ small regrets, and two *big* regrets:

I) We built Dapper to find patterns in traces, but we failed to make those findings *discoverable.*

II) We built Monarch for core monitoring, but we failed to make that monitoring *actionable.*

[Ben Sigelman @el_bhs]

3/ Dapper’s early-days usage issues boiled down to two core challenges:

a) The insights were *restricted to the tracing (span) telemetry*
b) Those insights could only be accessed *from Dapper.* (And hardly anybody “started” in Dapper from a UX standpoint)

[Ben Sigelman @el_bhs]

0/ This is a 🧵about my experiences building both the Distributed Tracing and Metrics infra at Google.

And, particularly, my regrets. :)

Here goes: 👇

[Austen Allred @Austen]

Lesson it took me too long to learn:

The worst thing you can do to a startup is distract it from what’s most important.

[Daniel Vassallo @dvassallo]

Small bets:

- Work with high intensity, for a short time, and not too often.

- Work on things that are very likely to pay off, even if the upside is small.

- Build on the easy wins to try some speculative bets with more upside.

- Do multiple things at the same time.

[Marc @marc_data]

Auto'magical instrumentation for the ppl.

[vicki @vboykis]

An EASY button, but for closing out all my social media tabs and Slack, and opening up my code editor on a blank desktop.