10,286 karma · joined December 28, 2011
hnchat:Jpx4bUmdBFA0Wx1DtALs
Just moved to NYC (blog hasn't been updated yet). Always interested in getting to know people.
Additionally, my personal experience with things like this is that human cashiers aren't any better than me at looking up produce in the absence of a PLU code. In fact, I'm generally better because I know what I picked up (or intended to pick up), so I just look for the right name.
That said, it's up to the individual company to decide if the added cost is worth it. Just because the cost exists doesn't mean it isn't worth it.
Second, that is not evidence in the slightest of people doing this "every few years". Hell, most people don't want to move locally that often, let alone an expensive and very disruptive long distance move.
I suspect that people who work for companies where teams, organizations, or even the entire company are colocated will tend to favor in-office work to some degree, while people who work for companies where teams and offices are geographically distributed will tend to favor remote work.
I personally worked for Twitter in-office full time for 14 months pre-COVID, and almost every meeting or collaboration I had involved a teleconference with people in another timezone or was done over Slack. There was some opportunity for face-to-face collaboration with people from other teams from time to time, but in my opinion not enough to justify the daily commute. That said, I'm open to a hybrid arrangement for companies where the situation I had at Twitter would not be the case.
Location: NYC metro
Remote: Yes (or hybrid)
Willing to relocate: No
Technologies: Scala, Java, Python, Hadoop, NumPy
Résumé/CV: https://wmkrug.com/krug-resume.pdf
Email: hnhiring@wmkrug.com
I can't really summarize myself as an "X engineer". My career path has not been very orthodox, but it centers primarily on compute-heavy software systems, generally high throughput and usually low latency. I have written code for desktop applications, a real-time processing pipeline, backend services, and batch processing jobs. My production code has used varying combinations of C, C++, Java, Scala, and various DSLs; I have also used Python heavily for data analysis and tooling. Non-professionally, I use Python quite a bit for data analysis and tools. I'm currently learning Go, and re-learning modern C++.I have been working in ML and NLP for nearly 11 years, primarily on productionizing models and building applications around them. My most recent stint is four years with Twitter, but that will officially end in a couple days. Here I've been able to build and optimize a variety of ML-based services used by several teams within Twitter as part of their own services. Note, though, that I am not a modeller. I know enough ML and NLP to understand how the systems should work and how the model integrates into it, but not enough to effectively build the models. Recently, though, I have started learning about the training and architecture of large language models, and I'm experimenting with training and evaluating smaller language models.
From looking at job reqs I'm interested in, it seems my biggest weakness right now is limited experience with cloud platforms. My only professional experience is at Twitter where I used some features of GCP in a non-production capacity (primarily BigQuery. The services I built and maintained ran in Twitter's datacenters, where we used Mesos + Aurora as an IaaS, Twitter-wrapped Memcached and Redis clusters, and various data stores for persistence. However, I am interested in developing production-level knowledge of public cloud systems. I have used Azure for some small personal projects, I'm currently learning the basics of setting up Kubernetes clusters on GCP, and I've adopted Docker for managing my personal server.
I'm looking to continue doing backend or infrastructure work on high-throughput systems, such as the type of ML infrastructure work I was doing at Twitter or the high-performance computing type work I did at Raytheon. While I have done some work in building web applications and can work on them when necessary, I don't really want to become a full-stack engineer.
No, they don't. Professional societies aren't unions. They don't police anything, governments do; the societies only provide input to the rulesmaking process for the profession. They also have no power to represent their members in employment-related matters like a union. They do provide some union-like services to their members, but it's the fringe stuff like continuing education to maintain licenses.
Also, these exist for software engineers (ACM and IEEE), but the industry (at least the Valley-centric part) largely ignores them as membership organizations and resist any attempt to produce professional regulations similar to what those other professions have.
Location: NYC metro
Remote: Yes (or hybrid)
Willing to relocate: No
Technologies: Scala, Java, Python, Hadoop, NumPy
Résumé/CV: https://wmkrug.com/krug-resume.pdf
Email: hnhiring@wmkrug.com
I can't really summarize myself as an "X engineer". My career path has not been very orthodox, but it centers primarily on compute-heavy software systems, generally high throughput and usually low latency. I have written code for desktop applications, a real-time processing pipeline, backend services, and batch processing jobs. My production code has used varying combinations of C, C++, Java, Scala, and various DSLs; I have also used Python heavily for data analysis and tooling. Non-professionally, I use Python quite a bit for data analysis and tools. I'm currently learning Go, and re-learning modern C++.I have been working in ML and NLP for nearly 11 years, primarily on productionizing models and building applications around them. My most recent stint is four years with Twitter, but that will officially end in a couple days. Here I've been able to build and optimize a variety of ML-based services used by several teams within Twitter as part of their own services. Note, though, that I am not a modeller. I know enough ML and NLP to understand how the systems should work and how the model integrates into it, but not enough to effectively build the models.
From looking at job reqs I'm interested in, it seems my biggest weakness right now is limited experience with cloud platforms. My only professional experience is at Twitter where I used some features of GCP in a non-production capacity (primarily BigQuery). However, I am interested in developing this knowledge. I have used Azure for some small personal projects, I'm currently learning the basics of setting up Kubernetes clusters on GCP, and I've adopted Docker for managing my personal server.
I'm looking to continue doing the type of ML infrastructure work I was doing at Twitter, or move back towards the high-performance computing type work I did at Raytheon. While I have done some work in building web applications and can work on them when necessary, I don't really want to become a full-stack engineer.
I agree with your assessment, but I do want to highlight that this condition is not rare for Twitter. Load is very spiky, sometimes during predictable periods (e.g., the World Cup, New Year's Eve) and sometimes during unpredictable periods (e.g., Queen Elizabeth II's death, the January 6th US Capitol attack). It isn't going to cause a total site failure (anymore), but it can degrade user experience in subtle or not-so-subtle ways.
An aside on the "anymore", there was a time when the entire site did go down due to high-traffic events. A lot of the complication in the infrastructure was built to add resiliency and scalability to the backend services to allow Twitter to handle these events more gracefully. That resiliency is going to help keep the services up even if maintenance is understaffed and behind a learning curve.
Speaking for the US, they do not. Joining a professional society is not a condition of licensure, nor is licensure required to join (I have been a member if IEEE since I joined in undergrad in 2001 and never even took the FE).
Continuing education requirements are set by the state you are licensed in. Texas (where I would have been if I had pursued licensure) requires a mere 15 Professional Development Hours per year, of which five may be self-directed. That is far, far below what our industry expects.
Also, I worked for four years as an electrical engineer without even being an EIT. Not all "real Engineers" require licenses.
> There is a software PE exam
Not anymore. Every state that once offered it has dropped it, and NCEES no longer maintains such a test.
> if that meant never needing to do another leetcode or take home assignment or whiteboard exercise in an interview cycle
Considering how low the requirements for maintaining a PE are after passing the initial exams, I don't think that has a snowball's chance in Hell of flying in this industry.
Ferrari and McLaren cars are a very exclusive club. Ferrari, to my understanding, explicitly limits production capacity to maintain that exclusivity. The practices of these two companies are in no way comparable to the practices of a large-volume automaker.
The vast majority of prison labor is for the direct benefit of the states or the federal government, not private companies. Still a huge mountain of perverse incentives, though.
> In 1990, about a dozen European countries had a wealth tax, but by 2019, all but four had eliminated the tax because of the difficulties and costs associated with both design and enforcement. Belgium, Norway, Spain, and Switzerland are the countries that raised revenue from net wealth taxes on individuals in 2019 with net wealth taxes accounting for 1.1% of overall tax revenues in Norway, 0.55% in Spain, and 3.6% in Switzerland for 2017.
The citation for those statements links back to the OECD, so they apparently don't count the Italian and Dutch taxes as "wealth taxes". The NPR transcript linked elsewhere says only three countries have it, so that interviewee may not be counting Belgium either (since the tax is solely on financial instruments and not total wealth).
The kind of mad person who tries to use a Hey.com email account with a listserv...
When I was in undergrad (granted, 20+ years ago) part of my aid was a couple thousand a year in direct cash payments to me. There were no restrictions whatsoever on this money.
I don't think that is the reason. When Raytheon or other contractors perform software work under a DOD contract (i.e., they charge the labor to a contract) the government generally gets certain exclusive rights to the software created. Raytheon is technically still the copyright holder, but effectively is required to grant the US government an irrevocable license to do whatever they want with the source in support of government missions if the code is delivered to the government. Depending on the contract, such code may also fall under blanket non-disclosure agreements. I believe both of these are incompatible with the GPL, and the latter with having a public fork at all.
The company could work this out with the government, but it would be an expensive and time-consuming process because government program offices are slow, bureaucratic, and hate dealing with small exceptions on large contracts. They might even still refuse to make the contract mods required at the end simply because they don't understand it or they are too risk averse. Legal is likely of the opinion that it isn't worth trying, and the Raytheon program office likely won't push them unless they can show a significant benefit for the company.
I agree, though, that doctors for non-humans are generally cheaper than those for humans.
Only by far-right nutcases.
> in most other developed countries he'd be considered center-right, or center at the minimum
I read this frequently here, but when I look at parties considered center-right I don't see much alignment with even the conservative wing of the Democratic Party. For instance, I do not think Angela Merkel and Joe Biden are roughly equivalent in positions; I think Merkel is measurably to the right.
Common for "internal" USG sites. I don't know if it's intentional.
> 2. If you try https://www.intelink.gov/, the browser immediately warns you that the site is not secure, because of certificate problems.
Internal USG sites use USG-generated root certificates and certificate chains. These need to be installed manually from USG sources.
> 3. If you still dare to venture ahead, you are greeted with this: "This is a United States Government computer system. This computer system, including all related equipment, networks, and network devices, including Internet access, are provided only for authorized U.S. Government use. U.S. Government computer systems may be monitored for all lawful purposes, including ensuring that their use is authorized, for management of the system, to facilitate protection against unauthorized access, and to verify security procedures, survivability, and operational security. Monitoring includes authorized attacks by authorized U.S. Government entities to test or verify the security of this system. During monitoring, information may be examined, recorded, copied, and used for authorized purposes. All information including personal information, placed on or sent over this system may be monitored."
The standard disclaimer on all internal and classified systems. I'm glad I no longer have to click through that daily.
How does outrage lead to more ad revenue? Who wants their brand advertised next to outrage? How many outraged people click on ads?