When I was in college, I was one of many students who accepted a somewhat prestigious federal scholarship to learn a “critical language” (a language not many others Americans learn). In return, I pledged to work for the federal government for at least a year. I was assured that this was not a downside but an opportunity - there were lots of great jobs in the State Department, DoD, USAID, etc, that would love to hire someone with my credentials. In fact, they even gave us a form of preference eligibility to make the process easier for us.
Ha! I spent over 6 months applying before I got a position. It took around 3 months for me to figure out that I had to essentially copy and paste the job description into my resume, and another 3 to figure out that I had to rate myself as highly as possible on the self-assessment. To this day, I’m still not sure what the key to cracking HR-led preliminary interviews was.
The job I finally got was not one that dealt with the foreign policy issues I was interested in, or used the language skills the government had paid for me to acquire, or even required a college degree. So I worked for the government for my required year, then left for law school.
This all took place in 2020. Just a few weeks ago (in 2024!!) I got an email from one of the many of positions I had applied to, telling me they were not moving forward with my application.
LLMs would be effective for matching job descriptions to actual experience, at scale (regardless of how the resume was written) without bias from hiring staff. There could be pushback on this but any change will have some.
Private companies that have tried this (though perhaps with less sophisticated tools) have received so much pushback that many have (at least publicly) abandoned the process.
Agree, it could be pretty tough. Maybe a pilot could be tried and then 3rd party evaluated to assess outcome fairness. Given Musk's interest in AI, maybe this is something that he could make happen.
I actually think "assessing outcome fairness" is one of the fundamental issues. As someone who has been interviewing and hiring engineers for over two decades something that I have always told my team is that "You do not have a responsibility to be fair to candidates. Your job is to hire excellent people for our team."
Sometimes those two goals overlap, but sometimes they do not. And in the cases where they do not it's important to understand what takes priority.
Yeah in terms of at least screening down to a level that humans can further investigate, this seems like an appropriate application. "Read all of these and forward the ones that have at least an 80% chance of being qualified."
I appreciate Jennifer's critique of hiring processes, which have become ineffective and more cumbersome the more we use false markers for equity and for fit. So much of the piece resonated with me. But the credibility of the piece was deeply undercut by omission of analysis on other levels. Nowhere in this piece does it address the fact that the Trump administration DOES NOT CARE about fair-- or qualified hiring, for that matter. There has been ample reporting on the absolutely disgraceful questions on the HHS questionnaire for new hires, created by RFK Jr. and his anti-vaccine, misogynistic partners. The Trump WH and cabinet teams are using *loyalty tests* for hiring. And the choices they've made so far in nearly every area are either nepotistic or tragically unqualified. When people are hiring Christian Nationalists and Nazis, sex traffickers, thieves, and con men... I mean... give me a break. You don't even touch on this.
Having worked in the FAANGs and now coaching government and seeing the stark difference in HR practices; it leads me to believe that HR practices in government are useless, disconnected, convoluted and need a complete overhaul; there needs to be deep involvement of the hiring manager in the process; much like corporate. Frankly government practices such as these are deeply ineffective.
Hope springs eternal, but I've heard noting from administration appointees about reform, just decreasing expenditure on staff. An impossible hiring process may HELP reduce staff, so why "fix" THAT?
I would be interested to know if there are government hiring practices that are currently working well at other levels of government or in other countries with which we could compare the federal process in the USA. The SME-QA sounds great for specialists, but might be tricky for hiring on potential or hiring generalists.
It's perhaps a small footnote compared to the structural problems you've identified here, but OPM should also work to incorporate more features of modern applicant tracking systems into systems like USAJOBS and Monster Government Solutions.
Instead, it seems like agency attempts to innovate in this space get shut down.
This piece resonated so hard with me.
When I was in college, I was one of many students who accepted a somewhat prestigious federal scholarship to learn a “critical language” (a language not many others Americans learn). In return, I pledged to work for the federal government for at least a year. I was assured that this was not a downside but an opportunity - there were lots of great jobs in the State Department, DoD, USAID, etc, that would love to hire someone with my credentials. In fact, they even gave us a form of preference eligibility to make the process easier for us.
Ha! I spent over 6 months applying before I got a position. It took around 3 months for me to figure out that I had to essentially copy and paste the job description into my resume, and another 3 to figure out that I had to rate myself as highly as possible on the self-assessment. To this day, I’m still not sure what the key to cracking HR-led preliminary interviews was.
The job I finally got was not one that dealt with the foreign policy issues I was interested in, or used the language skills the government had paid for me to acquire, or even required a college degree. So I worked for the government for my required year, then left for law school.
This all took place in 2020. Just a few weeks ago (in 2024!!) I got an email from one of the many of positions I had applied to, telling me they were not moving forward with my application.
LLMs would be effective for matching job descriptions to actual experience, at scale (regardless of how the resume was written) without bias from hiring staff. There could be pushback on this but any change will have some.
Private companies that have tried this (though perhaps with less sophisticated tools) have received so much pushback that many have (at least publicly) abandoned the process.
https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG/
I can only imagine the firestorm of criticism that would arise about bias if the US government took this approach.
To be clear, I'm not saying it would be a bad approach! I just think the hurdles to getting it accepted would be immense.
Agree, it could be pretty tough. Maybe a pilot could be tried and then 3rd party evaluated to assess outcome fairness. Given Musk's interest in AI, maybe this is something that he could make happen.
I actually think "assessing outcome fairness" is one of the fundamental issues. As someone who has been interviewing and hiring engineers for over two decades something that I have always told my team is that "You do not have a responsibility to be fair to candidates. Your job is to hire excellent people for our team."
Sometimes those two goals overlap, but sometimes they do not. And in the cases where they do not it's important to understand what takes priority.
Yeah in terms of at least screening down to a level that humans can further investigate, this seems like an appropriate application. "Read all of these and forward the ones that have at least an 80% chance of being qualified."
I appreciate Jennifer's critique of hiring processes, which have become ineffective and more cumbersome the more we use false markers for equity and for fit. So much of the piece resonated with me. But the credibility of the piece was deeply undercut by omission of analysis on other levels. Nowhere in this piece does it address the fact that the Trump administration DOES NOT CARE about fair-- or qualified hiring, for that matter. There has been ample reporting on the absolutely disgraceful questions on the HHS questionnaire for new hires, created by RFK Jr. and his anti-vaccine, misogynistic partners. The Trump WH and cabinet teams are using *loyalty tests* for hiring. And the choices they've made so far in nearly every area are either nepotistic or tragically unqualified. When people are hiring Christian Nationalists and Nazis, sex traffickers, thieves, and con men... I mean... give me a break. You don't even touch on this.
Having worked in the FAANGs and now coaching government and seeing the stark difference in HR practices; it leads me to believe that HR practices in government are useless, disconnected, convoluted and need a complete overhaul; there needs to be deep involvement of the hiring manager in the process; much like corporate. Frankly government practices such as these are deeply ineffective.
Jennifer-
Do you have any insight into which agencies or offices are using the SME-QA?
As always, thanks for your amazing work, and wishing you the best for the holidays.
Is there a compelling reason not to simply fire every person who works in HR at the Federal level and let department teams do their own hiring?
Hope springs eternal, but I've heard noting from administration appointees about reform, just decreasing expenditure on staff. An impossible hiring process may HELP reduce staff, so why "fix" THAT?
I would be interested to know if there are government hiring practices that are currently working well at other levels of government or in other countries with which we could compare the federal process in the USA. The SME-QA sounds great for specialists, but might be tricky for hiring on potential or hiring generalists.
It's perhaps a small footnote compared to the structural problems you've identified here, but OPM should also work to incorporate more features of modern applicant tracking systems into systems like USAJOBS and Monster Government Solutions.
Instead, it seems like agency attempts to innovate in this space get shut down.