doğa's Avatar

doğa

@dogaker

Revolutionary wanderer. Tirades about data science, tech, and metal. Principal Data Scientist at MSFT. Ex-Tweep. Alum of UPenn and McGill. Past regenerations include Doctor of Sociology. He/Him.

339
Followers
1,176
Following
159
Posts
09.06.2023
Joined
Posts Following

Latest posts by doğa @dogaker

First vacation in over two years when I left my work laptop at home. What could go wrong? 🤣

28.02.2025 11:33 👍 2 🔁 0 💬 0 📌 0

It's been a couple of weeks back to back when I'm trying to explain to people that having a PhD doesn't magically transform bad data into insights. If you haven't done your due diligence at design stage and collected garbage in lieu of experimental data, no suffixes can come to rescue.

24.01.2025 00:05 👍 2 🔁 0 💬 0 📌 0

Got Graspop 2025 tickets yesterday! Still need to figure out travel arrangements, but am super excited to make another trip to Belgium.

01.12.2024 12:18 👍 4 🔁 0 💬 0 📌 0
Post image

Beautiful day in Killington for Stifel World Cup!!! #GrandSlalom #Skiing

30.11.2024 17:23 👍 9 🔁 0 💬 0 📌 0

Black Friday is one of the busiest times in my company. Not a single issue from my team's pipelines came up or any alerts fired up to get anyone on call. I'm so proud of all the hard work my team's been doing and feels so good to see its results!

30.11.2024 13:18 👍 2 🔁 0 💬 0 📌 0

People often confuse alarmism with progress. You don't build a data product with endless strings of last minute all nighters. You build it through failing fast, iteration, and a clear vision.

26.11.2024 13:50 👍 0 🔁 0 💬 0 📌 0
Preview
How MNTN Is Bringing Hollywood Creativity to Small Business With MNTN, Deadpool star Ryan Reynolds and tech guru Mark Douglas pair innovative advertising technology with a hell of a lot of fun

MNTN CEO Mark Douglas and CCO Ryan Reynolds are on the cover of INC this month! Well deserved!

The article also talks quite a bit about MNTN Matched and Keywords targeting that I spent most of my past 14 months or so leading the 0->1 and GTM efforts in engineering. So cool!
trib.al/2MnMrVd

25.11.2024 14:51 👍 2 🔁 0 💬 0 📌 0

That's what I hear! It'll be my first time seeing them so looking forward to it!

24.11.2024 14:03 👍 0 🔁 0 💬 1 📌 0

Excited about seeing Pineapple Thief tonight and Zeal & Ardor tomorrow!

24.11.2024 13:14 👍 1 🔁 0 💬 1 📌 0

Just digging through The Last Will and Testament. §1 rips! Sounds like the last years of Opeth remaking Ghost of Perdition. I'll take it.

22.11.2024 11:42 👍 1 🔁 0 💬 0 📌 0
Post image

Also sounds about right...

20.11.2024 17:31 👍 2 🔁 0 💬 0 📌 0
Post image

There's still 40 days to go this year, but this seems about right.

20.11.2024 17:28 👍 0 🔁 0 💬 0 📌 0

That is when it all begins again which is the true iterative product development cycle of data science in real life.

20.11.2024 16:37 👍 0 🔁 0 💬 0 📌 0

Once you deliver your solution to great success and everyone owes you one everyone now has confidence in you and learned from your example with the sureness that they can now implement your ways into theirs and be even faster next time.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

That is the crucial moment to rise up to the occasion and deliver your value to build confidence with your stakeholders, which only is possible if you haven't been sitting on your hands while everyone was busy playing with shit on the wall.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

Too often it seems more business savvy to throw piles and piles of shit at the wall over a quarter instead of the sound data approach and for people to finally come ask for your way when they truly have burned your runway leaving you little leeway for your mistakes.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

When you can throw shit at the wall in a week, your estimates that suggest your solution will take about a month or quarter often makes people think that you're too academic and slow.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

Next you deal with the velocity issues because often people confuse an iterative process with throwing shit at the wall. Yes throwing shit at the wall is very quick, much quicker than what you can offer.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

And yes unlike a simple logic assumption that you hope works, a testable hypothesis can be validated.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

In fact the reason they need your services is the code complex problem they're facing cannot be solved with some smart hack, but with a methodical and meaningful model that needs to be designed, developed, and validated.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

They expect it should be just as quick as their failed attempt if not quicker through AI sorcery. So a lot of your valuable time is explaining to them you're not some omniscient superpower.

20.11.2024 16:37 👍 0 🔁 0 💬 1 📌 0

Being a data scientist is weird because stakeholders assume they already know all the answers to be attainable by using their own simple logic, which often fails. They often come to you when all those fail and expect you to implement your wizardry to their solution to make it work this time.

20.11.2024 16:37 👍 1 🔁 0 💬 1 📌 0

Iron Maiden and Slipknot at #GMM2025. Give me Heaven Shall Burn, Cult of Luna, Julie Christmas, and Cemetary Skyline in the bill and I'm so there!

20.11.2024 15:21 👍 0 🔁 0 💬 0 📌 0

Post a perfect album from the 90s that isn't Nirvana, Pearl Jam, Soundgarden, or Alice In Chains.

19.11.2024 15:36 👍 2 🔁 0 💬 0 📌 0

Any advice for a fun game that's optimized for PS5 Pro?

19.11.2024 14:34 👍 0 🔁 0 💬 0 📌 0

I'm not familiar with them. What are they like?

19.11.2024 12:49 👍 0 🔁 0 💬 0 📌 0

I kinda wish I hit skip instead of transfer from other consoles. I'll have fallen asleep by the time I start any games...

19.11.2024 03:43 👍 0 🔁 0 💬 0 📌 0

Is it just me or is the DualSense controller that comes with PS5 Pro a little heavier than the one that comes with PS5?

19.11.2024 02:48 👍 0 🔁 0 💬 1 📌 0

Say it with me, "Data products are built with data people!"

19.11.2024 02:47 👍 1 🔁 0 💬 0 📌 0

I agree. It's misleading to make any takeaways from just a handful arbitrary accounts with high variance in their production numbers.

18.11.2024 13:05 👍 1 🔁 0 💬 0 📌 0