Author Topic: ULA General Discussion Thread  (Read 135175 times)

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 4728
  • United Kingdom
  • Liked: 6857
  • Likes Given: 36
Re: ULA General Discussion Thread
« Reply #440 on: 11/22/2022 10:44 am »
ULA's software job postings include a "summary salary range (Colorado only)". I assume this is required by some Colorado law. The salary ranges of the postings I looked at ranged from $66k for the bottom of the range for a level 1 posting requiring 0-2 years experience post-bachelors to $188k for the top of the range for a level 4 posting requiring 6-10 years experience post-bachelors. The top payers in Silicon Valley are Alphabet and Facebook with median pay of about $290k (https://www.fastcompany.com/90757550/heres-what-big-tech-companies-paid-their-average-worker-last-year) so ULA isn't competitive unless ULA pays a lot more for more senior positions that they aren't hiring for right now. They won't get top talent this way but as I mentioned in my previous post that's probably OK given ULA's goals.

These aren't positions suited for silicon valley bros though. This is probably deep embedded/realtime operating system stuff, which is an entirely different kettle of fish. Considering this is touching potentially life critical stuff as well, the design and programming methodology is a very different world from "ship fast and break things" internet programming. Classic example is the shuttle OS programming crews, who were physically separated to prevent interaction, and overtime was basically forbidden. Admittedly that might be a bit extreme, but that was a serious dedication to code quality. If you had deep performance tuning SRE's looking for a new gig though, they might be suitable, but bringing the toolchains they depend on to the RTOS world might be a dealbreaker..
As a counterpoint: "ship fast and break things" has resulted in Dragon 2's software, where both the software itself and the systems engineering methodology behind it have been praised by NASA in terms of both quality and documentation. The traditional approach taken by Boeing has not paid similar dividends in code quality.

Online deadman1204

  • Full Member
  • ****
  • Posts: 1209
  • USA
  • Liked: 1044
  • Likes Given: 1524
Re: ULA General Discussion Thread
« Reply #441 on: 11/22/2022 03:04 pm »
ULA's software job postings include a "summary salary range (Colorado only)". I assume this is required by some Colorado law. The salary ranges of the postings I looked at ranged from $66k for the bottom of the range for a level 1 posting requiring 0-2 years experience post-bachelors to $188k for the top of the range for a level 4 posting requiring 6-10 years experience post-bachelors. The top payers in Silicon Valley are Alphabet and Facebook with median pay of about $290k (https://www.fastcompany.com/90757550/heres-what-big-tech-companies-paid-their-average-worker-last-year) so ULA isn't competitive unless ULA pays a lot more for more senior positions that they aren't hiring for right now. They won't get top talent this way but as I mentioned in my previous post that's probably OK given ULA's goals.

These aren't positions suited for silicon valley bros though. This is probably deep embedded/realtime operating system stuff, which is an entirely different kettle of fish. Considering this is touching potentially life critical stuff as well, the design and programming methodology is a very different world from "ship fast and break things" internet programming. Classic example is the shuttle OS programming crews, who were physically separated to prevent interaction, and overtime was basically forbidden. Admittedly that might be a bit extreme, but that was a serious dedication to code quality. If you had deep performance tuning SRE's looking for a new gig though, they might be suitable, but bringing the toolchains they depend on to the RTOS world might be a dealbreaker..
As a counterpoint: "ship fast and break things" has resulted in Dragon 2's software, where both the software itself and the systems engineering methodology behind it have been praised by NASA in terms of both quality and documentation. The traditional approach taken by Boeing has not paid similar dividends in code quality.
"Go fast and break things" is fine when you are developing a product, or have a product on the market that you can screwup and break (see game patches for example). However, spaceX is past this point with dragon and falcon.
If they screw up and "break things" causing a loss of mission, that will be enormous. So they don't. They carefully consider how the changes will affect things, test it to look for problems, ect. AKA move slow and be careful.

Its great for memes and random forum comments, but spaceX doesn't and cannot just "try stuff and see how it goes" with their established things like falcon 9 and dragon.

Online Robotbeat

  • Senior Member
  • *****
  • Posts: 36113
  • Minnesota
  • Liked: 20457
  • Likes Given: 10612
Re: ULA General Discussion Thread
« Reply #442 on: 11/22/2022 03:30 pm »
https://twitter.com/torybruno/status/1593640760464711680

Quote
Yes.  ULA is hiring.  All disciplines.  But especially software
And are the paying competitive wages compared to Silicon Valley to assure they get the software talent?
They basically just have to say they aren't twitter at this point.
Flight software is significantly different than software for a microblogging service. ULA employees with an expendable rocket are going to have to work pretty hard to stay competitive against the massive wave of competition from reusable rockets. I remember ULA requires pretty difficult hours, too, and they had pretty deep headcount cuts to enable being even their current level of competition. And the continued control of the company by Boeing has pushed out people like Sowers who wanted to pursue more innovation (depots, SMART reuse, etc).

ULA is very professional and does good work, but I’d be wary working for a company whose next-Gen vehicle doesn’t have a plan for a high level of reusability.

Flight software folks also don’t tend to make the same salaries as Silicon Valley engineers. Competition from reusable New Glenn is about 2 years out, so there’s not a long time of assured stability at the company, either.

Twitter software engineers can likely find better salary and conditions elsewhere unless they have a special interest in spaceflight. (Gotta say working at ULA is probably a better use of their talents than optimizing ad placement and toxic engagement…)
« Last Edit: 11/22/2022 07:14 pm by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline baldusi

  • Senior Member
  • *****
  • Posts: 8351
  • Buenos Aires, Argentina
  • Liked: 2530
  • Likes Given: 8047
Re: ULA General Discussion Thread
« Reply #443 on: 11/22/2022 05:41 pm »
[...]
"Go fast and break things" is fine when you are developing a product, or have a product on the market that you can screwup and break (see game patches for example). However, spaceX is past this point with dragon and falcon.
If they screw up and "break things" causing a loss of mission, that will be enormous. So they don't. They carefully consider how the changes will affect things, test it to look for problems, ect. AKA move slow and be careful.

Its great for memes and random forum comments, but spaceX doesn't and cannot just "try stuff and see how it goes" with their established things like falcon 9 and dragon.
That is a false dichotomy. SpaceX has perfected the high-reliability with high iteration model. It all depends on how you model your certification process. NASA took a long time accepting SpaceX practices, but now they are sold.
Traditional problem is when you spend most time writing requirements and certifying the design, making any change becomes exponentially expensive, even when it would actually lower your complexity and simplify verification in the end. Using an agile or similar process to actually write something, then learn from verifying that to make a better design, and iterate it many times gives a better product. You "just" have to do it thinking about verification and reliability as your main objective rather than features.
As always, good software management is taking a bootknife to requirements and features.

Offline su27k

  • Senior Member
  • *****
  • Posts: 6070
  • Liked: 8630
  • Likes Given: 840
Re: ULA General Discussion Thread
« Reply #444 on: 11/23/2022 08:28 am »
https://twitter.com/torybruno/status/1593640760464711680

Quote
Yes.  ULA is hiring.  All disciplines.  But especially software
And are the paying competitive wages compared to Silicon Valley to assure they get the software talent?
They basically just have to say they aren't twitter at this point.

LOL, so clueless about the current job market. Tech companies are laying off people left and right, Meta laid off 11k, Amazon 10k, both are bigger than twitter layoff. Google may be planning 10k layoff as well. And smaller companies are doing the same, Worldwide, more than 120,000 tech workers have lost jobs as a result of cutbacks by US tech companies, the whole twitter thing is not even a blip on the radar.

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 4728
  • United Kingdom
  • Liked: 6857
  • Likes Given: 36
Re: ULA General Discussion Thread
« Reply #445 on: 11/23/2022 11:49 am »
ULA's software job postings include a "summary salary range (Colorado only)". I assume this is required by some Colorado law. The salary ranges of the postings I looked at ranged from $66k for the bottom of the range for a level 1 posting requiring 0-2 years experience post-bachelors to $188k for the top of the range for a level 4 posting requiring 6-10 years experience post-bachelors. The top payers in Silicon Valley are Alphabet and Facebook with median pay of about $290k (https://www.fastcompany.com/90757550/heres-what-big-tech-companies-paid-their-average-worker-last-year) so ULA isn't competitive unless ULA pays a lot more for more senior positions that they aren't hiring for right now. They won't get top talent this way but as I mentioned in my previous post that's probably OK given ULA's goals.

These aren't positions suited for silicon valley bros though. This is probably deep embedded/realtime operating system stuff, which is an entirely different kettle of fish. Considering this is touching potentially life critical stuff as well, the design and programming methodology is a very different world from "ship fast and break things" internet programming. Classic example is the shuttle OS programming crews, who were physically separated to prevent interaction, and overtime was basically forbidden. Admittedly that might be a bit extreme, but that was a serious dedication to code quality. If you had deep performance tuning SRE's looking for a new gig though, they might be suitable, but bringing the toolchains they depend on to the RTOS world might be a dealbreaker..
As a counterpoint: "ship fast and break things" has resulted in Dragon 2's software, where both the software itself and the systems engineering methodology behind it have been praised by NASA in terms of both quality and documentation. The traditional approach taken by Boeing has not paid similar dividends in code quality.
"Go fast and break things" is fine when you are developing a product, or have a product on the market that you can screwup and break (see game patches for example). However, spaceX is past this point with dragon and falcon.
If they screw up and "break things" causing a loss of mission, that will be enormous. So they don't. They carefully consider how the changes will affect things, test it to look for problems, ect. AKA move slow and be careful.

Its great for memes and random forum comments, but spaceX doesn't and cannot just "try stuff and see how it goes" with their established things like falcon 9 and dragon.
"Try stuff and see if it works" is exactly what SpaceX do: because their software development lifecycle assumes full end-to-end system testing and simulation for all changes, with documentation of all changes.

My point was that rapid iteration or structured coding are not inherently better systems engineering methodologies than one another, you just have to do them properly. You can screw up waterfall and end up with patchwork incomplete tests and code full of bugs that reaches flight hardware, or you can iterate 5 times a day with 5 full end-to-end hardware-in-the-loop simulations a day and end up with solid code for flight, just as you can rigorously follow 'traditional' coding approaches, write everything in ADA, and have code that passes rigorous final testing and performs flawlessly in flight, or have a rapidly iterated pile of uncommented spaghetti code that passes unit tests but fails almost immediately on flight hardware.

Online Robotbeat

  • Senior Member
  • *****
  • Posts: 36113
  • Minnesota
  • Liked: 20457
  • Likes Given: 10612
Re: ULA General Discussion Thread
« Reply #446 on: 11/23/2022 02:13 pm »
SpaceX does more waterfall than people think, FWIW.
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Tags:
 

Advertisement NovaTech
Advertisement SkyTale Software GmbH
Advertisement Northrop Grumman
Advertisement
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0