We’ve seen what black-swan events like COVID-19 can do to businesses. If AI products are left unmonitored, they could be delivering bad decisions to customers. Here is a detailed blog on how to address this. #ExplainableAI #MLOps #Monitoring medium.com/fiddlerlabs/ex…

Training Data. The neglected in academia, but an important topic in industry. I learned the hard way- errors in training data can easily wipe out accuracy gain from a model. I'm writing codes to monitor data fed to ML. Things to check (can you add more?): 1. skewness between

One of the single biggest contributions of COVID-19 to businesses world-wide and society as a whole is that it has made everyone a data analyst and hungry for more insights. /thread

The Engineering Ladder (explained in self driving terms ) Level 1 Engineer - Hands On Wheel 1/4

To understand the territory of #MLops, first map out your #machinelearning workflows and ML systems: chapeau.freevariable.com/2020/10/mappin…

NO, your Machine Learning Models are NOT perfect! They need to be continuously monitored for garbage inputs, shifts in the input distribution, anomalies in feature data, performance decay, potential bias. A Reddit poll on Model Testing & Monitoring. reddit.com/r/MachineLearn…

Ok here we go. Look at the electoral maps by county for the last few decades of US presidential elections.  You’ll notice that the South goes almost uniformly Republican red every time. Duh. But if you look closer, there’s something else there ...

The Lt. Governor of PA is salty af this morning and I love it 🤣 “The president can sue a ham sandwich, it wouldn’t make a difference”

Our research has found that red & blue US political maps create increased perceptions of polarization and more political stereotyping, compared to more accurate purple maps. spssi.onlinelibrary.wiley.com/doi/abs/10.111… Thanks to Greg Albers for this new website: purplestatesofamerica.org

I kinda love Etsy

In 2021, what do you think is the single most important problem that the AI community should be working on?

Insightful thread from Krishna Gade around how to track ML model performance at scale.

Great resource for someone looking to catch-up on recent adavances in neural IR

Can cluster planting trees make California more wildfire-resilient? New science has re-sparked the long-lasting debate on how to best manage forests. [THREAD] bloom.bg/3xHD4Kn

A song from my teenage years is now finding relevance in my adulthood 😭

A great compendium of via distributed systems are hard at scale. Great write up and look back into leanings Yao Yue 岳峣 Dan Luu 🙌🏾

Vladimir Kostюkov @TwitterEng @finagle twemproxy was a life saver while adding Redis to Yahoo’s production stack many years back, saving us from having to deal with the quirks of each language’s immature Redis client libraries.

A look back in time. Some gems here :)

One of the things that I think is sad about the decimation of Twitter eng is that Twitter was doing a lot of interesting (and high ROI) engineering work that, at younger companies, is mostly outsourced to "the cloud" or open source projects A few examples off the top of my head:

My team at Google Deepmind in NYC is hiring research engineers! If you're interested in helping add new capabilities to large-scale generative models, please apply: boards.greenhouse.io/deepmind/jobs/… Feel free to reach out if you have questions :)