Manju Rajashekhar

EIR at Storm Ventures, Co-Founder & CTO of Blackbird AI - Acquired by Etsy, Ex-VP of Eng @Etsy, Early at @Twitter, @VMware, @Stanford, Avid 🏊🏾🧘🏾‍♂️

[Manju Rajashekhar @manju]

Thrilled to be joining the Etsy team! :) /cc @chaddickerson @mikebrittain @nikhil_raghavan https://t.co/PxkEChRLEI

[Manju Rajashekhar @manju]

@paulinebhyang Thank you for having me on your podcast! I appreciate all your insightful questions on such a timely and relevant topic. Loved the conversation & sharing our learnings on prototyping vs. deploying AI/ML at scale and deriving repeatable value from it.

[Pauline Yang @paulinebhyang]

Today’s guest on Cross Validated is @manju Rajashekhar, who is the VP of Engineering at @Etsy. Manju has spent over seven years at Etsy, was an early employee at @Twitter, and has spent time at @VMware and @Microsoft. Etsy is an ecommerce platform connecting millions of creative buyers and sellers around the world.

In this episode, we discuss:
- Mature use case of search, ads and recommendations on the Etsy platform
- Framework of deriving value from AI / ML
- Ease of prototyping vs. deploying AI today
- Strong belief why open source will win

Subscribe for the next episode: https://t.co/EPaddBrFNy

[billy bubba bingus-drumpfus @wgrathwohl]

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:

https://t.co/SRgn5ag79w

Feel free to reach out if you have questions :)

[Manju Rajashekhar @manju]

@ttunguz Congrats Tomasz! And best of luck on what's next. Looking forward to it

[Dan Luu @danluu]

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:

[Manju Rajashekhar @manju]

@thinkingfish What a great run, Yao! Thank you for everything you did <3

[Manju Rajashekhar @manju]

@normanmaurer Congratulations! So well deserved.

[EJ Campbell @ejc3]

@vkostyukov @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.

[Manju Rajashekhar @manju]

A look back in time. Some gems here :)

[Vladimir Kostюkov @vkostyukov]

A few takes on how hardcore @TwitterEng has been. Let's talk about @finagle and Scala for a minute.

[Manju Rajashekhar @manju]

@thinkingfish @jreichhold Another incident I recall was with LRU allocator in fall of 2010 which evicted hot keys leading to doubling of traffic to DB. This was the impetus to create Random allocator and the birth of Twemcache in preparation for our upcoming Super Bowl traffic

[Manju Rajashekhar @manju]

@thinkingfish @jreichhold Also recall the time when all the caches were restarted simultaneously instead of rolling restart 😬 and I got a call from @rgbenson at night with a message: “manju, don’t panic but all the caches have been restarted”. And I replied to rob as “you mean by all, ALL caches”. 😱

[Manju Rajashekhar @manju]

@thinkingfish @jreichhold If I recall there was a timeout bug in libmemcahed that would just hang the client that caused all our rails worked to hang

[Manju Rajashekhar @manju]

A great compendium of via distributed systems are hard at scale. Great write up and look back into leanings @thinkingfish @danluu 🙌🏾

[Dan Luu @danluu]

A history of major cache incidents at Twitter

https://t.co/Lp4biTUPDe

[Manju Rajashekhar @manju]

@arosien Congratulations, Adam <3

[Manju Rajashekhar @manju]

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

[Doug Turnbull @softwaredoug]

Dear fellow relevance engineers, allow me to karaoke you

🎶Don't go chasing state-of-the-art!

Please stick to the methods and the techniques you're used to

Take time to measure and evaluate

don't assume SOTA work well for youuouu... 🎶

[Bloomberg Green @climate]

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] https://t.co/SPxTcGiXJx

[Manju Rajashekhar @manju]

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

[Sebastian Hofstätter @s_hofstaetter]

Want to learn about some of the latest neural network advances in Information Retrieval🔎? Then join us for our online master-level course "Advanced IR" @tuvienna!
We'll revisit some IR basics, NLP techniques, and then focus on cool new neural IR methods🔥 https://t.co/Htk6RLaTVK

[Manju Rajashekhar @manju]

Insightful thread from @krishnagade around how to track ML model performance at scale.

[Krishna Gade @krishnagade]

Machine Learning models are the new software artifacts getting operationalized at scale today but they are hard to debug!

If I were an ML Engineer on-call and got a PagerDuty alert that CLICKs are down on our AI-based recommendations product, how do I troubleshoot it? /thread

[William Benton @willb]

To understand the territory of #MLops, first map out your #machinelearning workflows and ML systems: https://t.co/PSUgn09S4L

[ethn @ethanyish]

The Engineering Ladder (explained in self driving terms )

Level 1 Engineer - Hands On Wheel 1/4

[Andrew Ng @AndrewYNg]

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