DOGE engineers are working on AutoRIF, a software system connected to reductions in the federal workforce, according to sources cited by WIRED. The work is drawing attention because thousands of workers have already been terminated over the last few weeks across multiple agencies, and some employees fear automation could speed up the next phase.
What AutoRIF Is Built To Do
AutoRIF stands for Automated Reduction in Force. It was first developed by the Department of Defense more than two decades ago and has since been updated several times.
The software has been used by a variety of agencies to expedite reductions in workforce. In plain terms, its role is tied to the administrative process that determines which employees may be affected when an agency carries out a reduction in force.
Government HR officials conducting RIFs are required to create ranked lists of employees who may be subject to firings. A former government HR official told WIRED that AutoRIF does that automatically.
That automation does not remove every human step. The same former official said, "However, even with the use of any automated system, the OPM guidance says all data has to be confirmed manually and that employees (or their representative) are allowed to examine the registers."
It is not immediately clear if AutoRIF’s capabilities have been altered either by the Defense Department or DOGE.
What WIRED Says DOGE Is Doing
WIRED reviewed screenshots of internal databases showing that DOGE operatives have accessed AutoRIF and appear to be editing its code. A repository in the Office of Personnel Management’s enterprise GitHub system is titled “autorif.”
That repository appears in a space created specifically for the director’s office, where Musk associates have taken charge soon after Trump took office. Changes were made as recently as this weekend, according to WIRED.
Sources at OPM speculated that Musk-affiliated engineers could be using AutoRIF as built, building software on top of it, or using code from it. The distinction matters because the source article does not establish exactly what DOGE intends to deploy, only that the code has been accessed and appears to be under active work.
WIRED also reported that Riccardo Biasini, a former engineer at Tesla and a director at The Boring Company, has seemingly been tasked with pruning AutoRIF on GitHub. His name is attached to the repository in screenshots viewed by WIRED.
One file description authored by a user with Biasini’s username on GitHub says, "Remove obsolete versions of autorif." Biasini has also been listed as the main point of contact for the government-wide email system created by the Trump administration from within OPM to solicit resignation emails from federal workers.
OPM did not immediately respond to requests for comment from WIRED.
Why Federal Workers Are Concerned
Until now, federal agency firings have been carried out manually, sources told WIRED. HR officials have been combing through employee registries and lists provided by managers.
Probationary employees have been targeted first. The source describes them as workers who were recently hired, promoted, or otherwise changed roles, and says they lack certain civil service protections that would make them harder to fire.
The concern among some government employees is that software and AI could allow large-scale terminations to move even more quickly. AutoRIF is already associated with expediting reductions in workforce, and the broader use of automation could make the process feel less visible to workers who may be affected.
That concern is not only about speed. It is also about whether agency-level information and manager judgments are being used when decisions are made.
The Email Push Adds Another Layer
The AutoRIF work comes as DOGE seemingly prepares for its second major round of firings, according to WIRED. On Saturday evening, government workers received another email purportedly from OPM asking them to reply with what they accomplished in the last week.
The request asked employees to provide five bullet points explaining their top work achievements of the last week. Some agencies, including the FBI, asked employees not to respond to the message.
In a meeting with HR officials on Monday, OPM told agencies they could ignore the email. On Monday, NBC News reported that the information would be fed into an unspecified large language model that would assess whether an employee was necessary.
The source article does not say how that model would work, who would review its output, or whether it would be used directly in termination decisions. It does show why workers are paying close attention to both the email system and the AutoRIF code work: each points toward a more centralized and automated approach to evaluating federal employees.
Mission Critical Lists Were Not Enough At CDC
WIRED also reported an example from the Centers for Disease Control. Before the first round of probationary firings, CDC managers were asked to mark workers they considered “mission critical” and send that list up the chain of command.
A CDC source told WIRED that managers made a deliberate effort to identify probationary employees whose removal would affect the mission. The source said, "CDC went through a very, very deliberate effort to characterize our probationary employees as mission critical or not, and that way we could keep those that would have real impacts to the mission should they get terminated."
The same source said that work did not appear to shape the final outcome: "None of that was taken into account. They just sent us a list and said, ‘Terminate these employees effective immediately.’"
That account captures the central issue raised by the AutoRIF reporting. Automation can make a process faster, but federal workers and managers are questioning whether the process will still reflect local knowledge, manual verification, and the consequences for agency missions.