Irene Y. Zhang

University of Washington Sexual Harassment and Bullying Investigation

For the last 6 months, the UW Allen School of CSE has been investigating the toxic and sexist environment in the systems and networking labs. Nine doctoral and postdoctoral women were in the lab or joined during the investigation; out of these, two left academia and two are now at other schools. The others remain at UW but avoid the lab by working from MSR or other labs. All have lost up to a year or more of their PhDs to harassment. Despite this impact, no one has faced any consequences or repercussions.

Instead, the university has chosen to blame me for the situation, gaslight the women, and continue to fail to provide a safe learning environment. This outcome is not surprising (see: DARVO); what is surprising is how many times the Allen School administration could have de-escalated the situation for the mutual benefit of the women and the school but instead let the men manipulate them and use the system to further harass the women. The remainder of this post is about my personal experience with the situation (usual disclaimers apply).

I’d like to thank my management chain at MSR for being supportive while I was involved in this situation. Generally, Microsoft seemed to take the stance that working with students is part of my job, and I should be able to do my job without experiencing inappropriate behavior. As a result, they also took it very seriously when my husband received threatening emails about the situation. I expect to get more after this blog post.

Overview

On July 12, 2024, seven women sent a letter to the systems and networking faculty (Tom Anderson, Baris Kasikci, Simon Peter, Ratul Mahajan and Arvind Krishnamurthy) detailing three harassment incidents and asking for, among other things, a lab code of conduct and sexual harassment training. They wrote this letter at the request of the grad counselors, after the men had reported them repeatedly to the advisors for making the lab an “unwelcoming” space by talking about “women’s issues”, including women’s access to medical care.

UW hired outside legal counsel to conduct a 5 month investigation, which resulted in this publicly available report. While the report finds that none of the incidents involving the three men rise to the level of “harassment” by a narrow legal definition, it would be difficult to imagine that the incidents described would be allowed at any company. For example, several of the incidents detailed in the report are explicitly banned by Microsoft’s sexual harassment policy.

Almost 9 months after sending their letter to the advisors, most of the women have still not rejoined their labs or returned to lab activities. The Allen School is now finally working on sexual harassment training and a code of conduct.

My Experience

I became involved in the situation around the time of OSDI 2024 because the harassers were following the women around the conference. I only became involved because:

  1. The situation did not seem serious enough to require much of my time. It was inexplicable to me why the advisors were tolerating the kind of behavior discussed in the report.

  2. I personally knew many of the faculty. I have known Simon, who advises one of the men, since 2008 and I have known Baris since 2020 when I chaired EuroDW with him. I’ve also known and worked with Tom since 2012 and Arvind was one of my PhD advisors.

  3. The women had already attempted to talk to the faculty for months and even asked another female faculty member to talk to the faculty in the lab (as there were no women faculty in systems or networking at the time).

The seven women that wrote the July 12 letter clearly had no effective advocate among the faculty. One of them had already dropped out of the PhD program, and I was worried more of them would quit. My primary goal was to protect the women from further harm and de-escalate the situation, while repairing the relationship between the women and their advisors.

Edited to add: I’ve posted a perspective from one of the women on my involvement.

After the HR investigation started, the faculty and students were instructed not to speak about the situation. Thus, I became the only person that could support the women and speak out. I only chose to speak publicly when:

  1. I was forced to explain the situation and support a new woman joining the lab. No one at UW explained to her why there were no women in the systems lab or that it might be unsafe for her to be around the men under investigation.

  2. I heard another credible report of harassment from the same men of a woman in another lab, as well as, concerning incidents of physical violence and intimidation (which are detailed in the report).

  3. Magda Balazinska (the school director) refused to implement any changes unless the women made a formal complaint to Title IX. As the harm perpetrated by university Title IX offices is well-documented (and the women sent an email with the research on that to the administration), I expected the reporting process would be traumatic and not let the women rejoin their lab any time soon.

These three things happened in quick succession about one month after the women sent the letter to faculty. I had thought the issues were limited to the systems lab and would be fixed quickly, but it became clear to me that the issues were more severe and pervasive than I thought. I chose to resign my affiliate faculty position and posted to Mastodon, Twitter, Facebook and the uw-systems mailing list to ensure that other women did not find themselves in the same position.

HR Investigation Outcome & Subjects

UW HR hired a lawyer from Inslee Best to conduct the investigation. The lawyer spent 2 months interviewing everyone involved, and then another 3 months to write the report. The investigation concluded that none of the behavior rose to the legal definition of harassment. However, the report has a clear bias towards protecting the university, leaves out facts that do not favor their conclusion, is sexist and also includes a number of inaccuracies. I did not have the mental or emotional energy to write a critique of the report, so produced a response from ChatGPT instead.

Edited to add: The university administration has asked me to remove the men’s names, but as I believe there is no legal basis for them to censor me, I have left them here.

Since the university will not hold the men accountable for their behavior, I name them here: Pratyush Patel (Iota in the HR report), Tapan Chugh (Theta) and Anish Nyayachavadi (Eta). Per the HR investigation, most of the bullying came from Anish, while the incidents of physical violence were Tapan, and Pratyush was emotionally abusing and stealing ideas and work from women. Since many of the events in the report did not happen on campus or happened between “friends”, they were not counted as harassment. Pratyush’s academic misconduct was considered to be outside the scope of the report. There are more details and evidence (e.g., slack messages, emails, text messages) in the report. The women also provided the investigator with many more emails and text that were not included in the report it did not support the investigator’s narrative.

Edited to add: By request, I have produced a ChatGPT summary of the report and the women’s timeline to help people grasp the situation.

I feel that it is important for me to disclose the men’s identity because, while the women are easily identifiable in the report, the men are not because there are vastly more men than women in the lab. Furthermore, as the men will continue to be in our communities, I do not want women who are not privileged enough to be part of the whisper network to suffer harassment from them. I am inspired by women that named names, including Dana Moshkovitz and the women that spoke out about the University of Michigan case.

I also stress that their behavior was enabled by their advisors, especially Ratul Mahajan (Omicron in the HR report) who advises or works with all of the men. Again, I want to ensure that all women can make informed descisions about grad school and advisors and not just those who have the privilege of knowing who is problematic and who is not. It is clear from their actions even now that the advisors prioritize the men’s careers over the women that left, and over the mental health and careers of the women that remain.

Lessons Learned

If you experience a similar situation in your department, do not make the same mistakes as the Allen School. While by no means comprehensive, some easy things for other CS departments to do:

  1. Make expectations of appropriate behavior clear. If inappropriate behavior pops up, deal with it swiftly and aggressively. It is much less painful to prevent these issues than to deal with them afterwards.

  2. Expect harassment when advising women. Statistically, all women in grad school will experience harassment; be prepared to spot it and help deal with it. Bystander training can educate you on how to effectively engage in these situations.

  3. Be aware (if you are not already) that university processes protect the university and not harassed students. As shown in Nicole Bedera’s book about Title IX, this priority often translates to protecting the harassers. If at all possible, resolve the situation without university administration involvement.

  4. Find an advocate that is not in a position of power over the student. Students will generally not want to talk to administrators or faculty unless they know they can be trusted. Most people are not explicitly trained to intake harassment reports, so the reporting process itself can often compound trauma. Having a trusted advocate to navigate the system is crucial to making sure that the victim is heard.

  5. Find peer support. The only reason any women remain is because there were multiple women to carry the emotional and mental load for the last year. All of the women in this situation wanted to quit at some point because the process is designed to wear down victims until they stop complaining and only kept going because one of the other women would take up the work engaging with faculty and admin.

  6. Rebuild trust after the situation is resolved. Institutional betrayal is where people are hurt by institutions and authority figures that are supposed to protect them and takes time to process. The lost trust may never be rebuilt, but restorative justice techniques may help.

It basically boils down to this: do not recruit women and other URM to your lab if you have no plan for how to keep them safe. There are plenty of experts and researchers who study these things (e.g., Nicole Bedera), read their books and educate yourself before you become a cautionary tale. This podcast is a great starting point.

Contact