brd

lemongrasscloud.com | Senior Kafka Engineer | Hybrid (Chicago)

Lemongrass is a SAP focused consultancy support customers as they look to keep up with SAP's "migration to the cloud".

We not so infrequently run into Frankenstein solutions built by our customers that need to be understood, cleaned up, and then potentially right sized or retired. If you enjoy and/or have experience rolling up your sleeves, diving into a poorly documented Kafka deployments, and coming out the other side the hero with crisp documentation in hand and a plan for longer term re-architecture activity, I've got the role for you.

If interested I'd love to talk to you, reach out to bdennett at lemongrasscloud.com

I'm reading a lot of comments here that are defensive about their phone usage. I think that misses the point. It's fine to chase productivity (real or otherwise) and we can all rationalize how we burn our spare minutes: decompression, etc.

Being able to be bored and have those creative thoughts enter is for sure a useful thing. I'd argue the more important thing is being comfortable with silence/boredom. Being able to sit in a meeting and let awkward silence stew; or make a sales pitch and quietly let the gears turn, is a super power. If you're the one at the table better with silence, you have an inherent advantage.

Context: I'd put myself in the very low phone usage category but I still use my phone far more than I'd like. I pretty much just check email/text, HN, a bit of news, and occasionally doom scroll reddit. I'm also a developer turned exec/sales guy.

I won't even bother clicking this click bait title. I learned Pascal as a sophomore in high school and by my senior year I knew I'd never touch it again. At the time it was a great introductory language but there's very little reason anyone should even be thinking about Pascal these days.

Fun fact: I still have my old pascal files on floppy disks in a box somewhere with a not so thin layer of dust on them.

MTurk has had this issue for years. If you tried to us MTurk to do any text or image labeling since at least 2018 you were majority of the time getting output from a poorly functioning ML system.

No, we've got our own fine tuning pipeline and initial tests showed better performance without traditional stem/lem processing so we dropped it from our classification pipelines and haven't seen a need to revisit.

I really appreciate how accessible SpaCy has made NLP work but their NER is definitely low accuracy.

Where stem/lem felt critical to successful NLP processing a few years ago, we've found stem/lem work to be much less important for downstream tasks when transformer based models are involved.

For topic extraction stem/lem still seems to do a lot to improve accuracy and for rules based approaches I can still see how it would facilitate more efficient processing at scale. I'd be curious to hear your experience fine tuning and/or training new models after stem/lem processing with transformers, we've admittedly done little testing to see how transformers actually performer if properly tuned to post-processed data.

I'm not at all surprised to see this. The writing has been on the wall that Hitachi has an appetite for getting into enterprise software. 5-ish years ago they bought an SAP infrastructure shop that I used to work with when I consulted in the space. At the time it seemed they were dipping their toes but it felt like only a matter of time until they made a bigger push into enterprise software services.

This makes me uneasy. The fact that there will be regulatory investigations should also concern you since a solution like this could easily be implicated in misleading retail investors.

Simple string matching isn't going to cut it, you really need a decent NLP based implementation. There should be weighing mechanisms to avoid bots surfacing stocks that don't actually represent the interest of WSB. The site should be crystal clear about the time period this is sourced from. And most importantly, please put a disclaimer about what it is, how it's working, and why it shouldn't be considered financial advice.

Amusingly enough, I've got a document that's essentially a list of ideas where the business model is spurious at best and on that list is a solution almost identical to what was built here.

I appreciate that someone took the dive and then shared the process, pain and failure of seeing it through to its conclusion. The only thing missing is the attempt to raise VC money to scale it to some sort of expert system you can sell to enterprise customers.

As an industry we're letting history repeat itself and making all the same mistakes.

There are different kinds of developers. At it's most base form, you have systems focused developers and algorithmic focused developers. Sure there is a grey area but I think those two buckets are pretty defensible.

In the data science world you have an exact parallel. Those who build the systems and those who optimize the thing the system supports.

In the ML world you have another parallel. Those who build the systems and those who optimize and pioneer the model architectures and parameters.

We never reached consensus on the titles for different kinds of developer/programmer/computer scientists. And we're failing now to reach consensus on sane titles for ML and DS.

My guess is theres some sort of yet to be identified global issue. Seems Comcast has been spotty in parts of the US, slack is down, youtube was briefly inaccessible for me, etc.

The US in particular has taken the pursuit of convenience to extremes and the abuse of plastics is one of the clearest examples of that. Most packaging should be glass, paper, or metal. Fewer things should be so disposable.

One thing I wish we could find a solution to is our oil subsidy problem. I get that the US gov wants to subsidize oil for the sake of national security but those subsidies have all sorts of crazy unintended consequences. One of those consequences was that oil found a way to make itself the primary component of packaging.

Having consulted for a major glass manufacturer I can assure you that recycled glass is a coveted resource. It's easier to work with and it's less energy intensive. I wasn't involved much in that side of the business so I don't know the details but I was given tours and cullet (recycled) glass was a BIG deal.