Vijay Chidambaram
@vj_chidambaram
Followers
10K
Following
101K
Media
255
Statuses
24K
Prof, UT Austin CS Dept. Views my own.
Austin, TX
Joined June 2012
Started work on the book! . Title: "CS Assistant Professor Handbook".Website: If you would like to get updates, please sign up here! (and let me know what else you would like to see in the book!)
🚨 Planning to write a smallish book: "Things I wish I knew when starting as an Assistant Professor". 🚨. Planning to cover things like :.- how to find good students looking past GRE/grades.- how to get funding from industry.- more below. What would you like to see included?.
3
36
217
I got tenure! Associate prof starting from Fall 22. Thanks to my wonderful students at @UTSASLab and everyone who supported me during this time!
138
10
1K
The number of African Americans granted PhDs in CS in 2019 in the entire country (USA) was. 13. 10 male, 3 female. I knew the number was low, but I was shocked to find this out!. Source: Taulbee Survey Thanks to @blwinks for sending me this reference!.
29
347
903
So the stated reason for censoring @timnitGebru's paper (and subsequently leading to her dismissal from Google) seems to be that it ignored relevant research, and this is pretty much BS. This is the job of conf reviewers, and not the job of Google.
15
119
734
List of universities not requiring the GRE for computer science PhD admissions this year!.
UC Berkeley.MIT.Stanford.CMU.UIUC.U. of Washington.Cornell.Georgia Tech.Princeton.UT Austin.Michigan.Wisconsin.UCSD.Harvard.UMD.UPenn.Purdue.UMass Amherst.NYU.NEU.UChicago. And I hear more are coming. I'll add more if people reply.
13
150
639
PhD/MS applications writing SOP/research statements, here is a series of tweets on what to avoid, and what to mention in your SOP. Targeted especially at international students. 1/n.
16
263
612
Okay, another question via DM: how to decide between academia versus industry job versus post-doc? A 🧵. There are a number of differences between academia and an industry job, but to me it boils down to three things:.1) Students.2) What you want to work on.3) Money.
13
86
484
So my students are off for summer research internships, and I gave them some advice. This is widely applicable, so sharing it here. First off, the worst way you can spend your internship is to focus exclusively on the internship project and talk only with your mentor all summer.
9
84
450
This is false and harmful. It projects “good” programmers as people who do it for fun: competence has nothing to do with this, and plenty of programmers with families see it as a job, and they are excellent at what they do.
‘Most good programmers do programming not because they expect to get paid or get adulation by the public, but because it is fun to program.’ - Linus Torvalds. #tech #education #programming
22
96
419
Students who submitted your grad school application: it is *totally okay* if you read your application and find a few typos that you missed while proof-reading. Your admission decision will not hinge on typos. Don't worry about it.
7
24
393
My perspective on accepting papers changed during SOSP 2017, when I got to chatting with this senior professor who is extremely well-regarded in the field. We were chatting and he challenged me "What harm does it really do if a conf accepts a weak but technically correct paper?".
21
80
376
Being a PhD student is a job. The student becomes an apprentice for 3-6 years, and learns by doing research. The student is paid a stipend in exchange for doing research. This is different from a BS or a MS degree, where the student pays tuition fees to the university.
What‘s something that seems obvious within your profession, but the general public seems to misunderstand?.
17
69
341
Congratulations to Dr Jayashree Mohan @jayashree2912 on passing her thesis defense today :) The first student I independently advised at @UTCompSci :) Super proud of her work! . Jayashree will be joining @IndiaMSR after her graduation.
23
4
328
A bit of personal news: my wife delivered a baby boy recently :) Both mom and baby are doing fine! . Becoming a parent is exciting and overwhelming :).
28
1
325
Students, not getting accepted to PhD programs doesn’t say anything about you: I got rejected from the university where I’m now a professor and I have plenty of prof friends who all got many rejections before something worked out. Have faith in yourselves!.
Just wanted to throw lots of love to all the students who didn't get accepted into any PhD programs this year. My first round of PhD applications I got rejected from all 11 schools I applied to. Some days it's so draining feeling genuinely happy and excited for your peers (1/3).
1
35
310
This makes me sad, because I always thought of Google as a nice place to work and to do research. If you can't publish work that the undisclosed Gods at Google don't approve of, it is not worth doing research there.
2
23
260
My fellow profs, when someone asks you for a letter for a job application, and you take that opportunity to write a negative letter, it reflects badly on you, not the applicant. I can't believe people actually do this!.
11
26
268
I have a colleague who is amazing at saying no and setting boundaries. "Can you do this [totally non-urgent] committee work over the weekend?". "Sorry, I've plans to play with my child the whole weekend". It hadn't even occurred to me to say no - I'm glad they are my colleague!.
2
8
253
A quick thread on why academics (especially students) should be active on Twitter. 1/n. The short version: Twitter is like a global pub where fellow academics all hang out. You get both company for yourself and visibility for your work. You also learn about cool work from others!.
4
52
256
@timnitGebru If I wrote a paper at a company, and undisclosed folks wanted me to retract it because they felt the scholarship wasn't good enough, I would be pretty annoyed too. The formal review process is for making sure that the paper isn't leaking company secrets etc.
3
13
228
I finished high school in India in 2009. If you track my cohort today 12+ years later, you would find it very hard to figure out who did well in school and who did not. We dramatically overestimate the impact of grades on success in life (i say this as an educator).
9
15
242
🚨 Planning to write a smallish book: "Things I wish I knew when starting as an Assistant Professor". 🚨. Planning to cover things like :.- how to find good students looking past GRE/grades.- how to get funding from industry.- more below. What would you like to see included?.
21
7
230
Congratulations to everyone landing CS faculty offers!. Please check out my book, "The CS Assistant Professor Handbook": I talk about things like funding, teaching classes, recruiting students, and so on. I've gotten feedback it is useful for new faculty!.
1
44
226
What I would love to see happen:. Universities band together to say: asking international students to take in-person classes just to stay in the US is a terrible idea, and we will not do it. It imperils our students needlessly. We will challenge this in court.
6
26
214
All the technically strongest people I know are *kind* people. My advisors/profs at @WisconsinCS, my colleagues at @UTCompSci, they are all competent, caring, empathetic human beings. Sure, there are some jerks, but they are the minority -- there is no need to hire them.
5
20
220
It has begun! Margo Seltzer moves from Harvard to UBC. My prediction: this is only the first in a long line of researchers who will be moving from the USA to Canada for better funding (Margo gets $1M each for 7 years), and more inclusive/non-toxic immigration policies.
I'm very excited that Margo is joining @UBC_CS !
7
61
185
Hearing about profs who want 10 hours in the lab 6 days a week AND weekend meetings, so for graduate students who encounter such profs:. This is not normal. This person is not a good PhD advisor. Do not sign up to work with them even for one day if you hear anything like this.
6
31
200
Just released the first draft of the Funding chapter in the "CS Assistant Prof Book". With an example budget, and answers to common questions such as "what are indirect costs"! . Book at:
6
30
198
Many papers have great stories behind them. It is a pity that we publish the paper and the story is lost! . The stories are often inspirational, and help junior grad students see the struggle behind influential work.
7
40
191
So frustrating when a systems paper makes big, complex advances to the state-of-the-art and then keeps the code private. You can't build on the code, you can't run new experiments to compare against them, nothing. Publishing for the sake of publishing.
4
18
197
Our work on the SquirrelFS file system in Rust got accepted to #OSDI24! This is work led by @hlebland, joint with @ntalyour and @siderealed. We use the Rust type system to check for crash consistency at *compile time*, eliminating ordering-based crash-consistency bugs.
7
22
192
Update as of 8th night: MIT, Harvard, and UC are suing, while Cornell, Princeton, Dartmouth, USC, Stanford, and Purdue are filing briefs of amicus curiae!. Heartening to see so many universities support international students!.
What I would love to see happen:. Universities band together to say: asking international students to take in-person classes just to stay in the US is a terrible idea, and we will not do it. It imperils our students needlessly. We will challenge this in court.
4
24
191
I wish research has the culture of "fostering": I'll send over one of my students to your lab for a semester/year, and I'll mentor your students for the same time. Would give students a lot of perspective and allow good research mgmt ideas to spread beyond one lab!.
14
10
193
I teach this paper (and the preceding Disco paper) in my undergrad Virtualization class. The students love hearing about the tale of how VMware and ec2 came to be, and how it all started with academic research. It's a great ad for systems research!.
Was "Xen and the Art of Virtualization" the most valuable academic CS paper of all time?. While virtualization was already common on high-end servers, this showed how to on cheap x86 boxes. It lead directly to AWS, VMWare and Azure -- likely trillions of dollars in revenue.
3
17
190
Different people get into academia for different reasons, but I got into it because I love working with and nurturing students. Someone once said, “as professors, our products are the students. Publications are simply the mechanism by which we train students”.
5
8
186
So the next time graduate admissions come along, I will have to try and convince bright international female scholars to likely move to a country where they will have fewer rights and protections than their home country. Can imagine what an attractive proposition that will be.
4
11
173
A lot of this resonates with me, and it's been hard to do any work for the past week or so.
3
16
184
Excited to share that our paper on understanding the impact of GPDR on database systems has been accepted to @pvldb! Excellent work led by @SuprShastri (who is on the job market) in collab with @TweetAtAKK, @VinaySBanakar, and @MelissaWasserma at UT Law!.
2
22
171
Chubby paper: "We claim no new algorithms or techniques. The purpose of this paper is to describe what we did and why, rather than to advocate it". Chubby won the SIGOPS Hall of Fame Award in 2017. Think of that next time you deride a paper as "just engineering".
5
42
162
Never thought I'd see the day when a US VP candidate is part Tamil :) And not n-generations ago Tamil, but to actually have a mom born in Chennai and speaking Tamil to this day! . Given that US presidents and VPs have to be born in the US, this is as close as you can get!.
2
7
169
Note that I'm basing this off on the email from Google folks:
6
9
153
So much snow in Austin! Got confused whether I woke up back in Madison ❄️
9
0
165
There seems to be enough interest in this that I'm starting a blog: Here's a first post describing what I'm planning to do: If you have interesting stories about your research projects, please send them my way!
Many papers have great stories behind them. It is a pity that we publish the paper and the story is lost! . The stories are often inspirational, and help junior grad students see the struggle behind influential work.
4
35
156
We found a bug in a verified file system! We ran Crashmonkey's suite of tests on MIT's FSCQ and found that it does not persist data on fdatasync properly. We emailed the authors, they have acked and fixed the bug. Come see our paper at #osdi18! . Details:
Excited to share our #osdi18 paper on finding crash-consistency bugs in Linux file systems! I will explain the intuition behind our system in this thread. Paper: Code : . @jayashree2912 made a demo video:
4
24
151
Given the randomness in our review decisions, even good papers can get multiple rejects in a row. This can be demoralizing and potentially lead to good people leaving the field. All the talk of "be persistent" is nonsense: what we should do instead is minimize random rejects.
2
10
149
Happy to share what I'm doing for my sabbatical: .- Part of Summer 2022 and Fall 2022 in India working with @vmwareresearch! .- Spring 2023 at @EPFL_en hosted by @JamesLarus . Looking forward to making new connections at both places! Please reach out if you would like to chat!.
6
2
154
So @MSFTResearch India has this excellent Research Fellow program where folks after undergrad can work with researchers for 2 years before applying to grad school. I call this a "pre-doc". Are there other pre-doc programs? Why aren't they more common?.
15
16
154
Maybe I should collect a thread of some of the horrible things that are happening in academia and show the thread to anyone who wants to apply for grad school. The positive stuff gets a lot of publicity, but there is a lot of toxic stuff too.
4
25
149
Hi folks, @NeerajaJY, @PooyanJamshidi, Romain Jacob, and I have a proposal about a new diamond open-access Journal for Systems Research. Concrete proposal: Blog post: Take a look and give us your feedback!.
7
48
151
When I was younger I read a lot of life advice and time management books. One thing that stayed with me: classify tasks as important and urgent, important but not urgent, urgent but not important. To grow, you need to do important but not urgent tasks every month.
2
22
143
Folks, you DON'T need to ignore your family and friends, or your health, to do good research.
2
19
143
It's so bizarre seeing @timnitGebru get commendation after commendation on her work, while at the same time get harassed by numerous Twitter accounts and gaslighted by Google.
Top read last month, how @black_in_ai, @AJLUnited and @Data4BlackLives are framing the conversation on equity in AI. 🧠🤖. Special commendation for @timnitGebru and her work in the field!. via @FastCompany
0
26
132
Here is something I wish as a community we understood and adopted in systems: . Novelty can come in many forms. Sometimes novelty lies in an interesting solution, sometimes novelty lies in an interesting new problem. Applying known techniques to a new problem is NOVEL.
2
19
145
The reason we dont do this, of course, is prestige. An acceptance rate of higher than 20 (25%) is scandalous. Shouldn't the prestige come from paper impact rather than acceptance rate? . The trouble is, impact takes a while, and we want something right now. .
5
9
139
This is insane. Every STEM grad student I know used OPT to work while waiting for a H1B visa. If this passes, expect number of international applicants for masters and PhD to go down drastically. International masters students pay high fees, crucial to keep these programs running.
The OPT program is crucial for retaining talented international students in the US. I relied on the OPT myself for summer internships during college and for full-time work after graduation.
1
22
137
This is disturbing. A PhD student from the University of Florida hanged himself because he felt his advisor was engaging in academic misconduct and he couldn’t take the pressure. He had a paper accepted to ISCA 2019.
So @Medium has published documents describing the scientific fraud and abusive lab culture that allegedly lead Huixing Chen to take his own life. Trigger Warning: (1/n).
5
85
140
Excited to share this Youtube playlist of videos we made @UTCompSci and @LAITSatUT where I explain/discuss basics in operating systems! Its loosely based on @WisconsinCS Three Easy Pieces OS textbook. Playlist includes demos and code walkthroughs! .
3
39
140
I've been a professor for 5 years now and I still get very anxious when looking at reviews. So if you are a grad student getting anxious on getting reviews, its totally okay! . I don't read the reviews immediately after getting them -- I have to "prepare" for it mentally.
2
7
139
Wow. Finally some news about this! Evidence found for everything alleged by the grad student who committed suicide. I hope there are heavy penalities all around, and not just slaps on wrists.
Results of the ISCA reviewer misconduct investigation are now public. Couple of highlight screencaps here because there wasn't a link to a press release by email and it's not up on SIGARCH yet
4
25
138
Students from India often ask me this question: "I did my undergrad at a small place with no research. I don't have great grades. How do I get into research and go to grad school?" . Ans: Work with a good prof (remotely if needed) and get their letter of recommendation.
3
28
138
I think there is strong demand for a new computing organization to replace ACM that does the following:.- Has open access as a core principle.- Is truly international, with both representation and decisions reflecting this.- Is centered around students.- Is agile and responsive.
8
13
134
Why isn't there a tool that takes as input:.- latex file .- a page limit. Output: modified latex file with vspace and other space hacks as needed. Will save innumerable hours of space hacking for researchers!.
11
10
135
I just reviewed a bunch of OSDI papers, a lot of them will get rejected due to either writing or evaluation. So here's a quick thread on how to do good evaluation in a systems paper. Others have said this before, but I think it bears repeating.
4
40
130
It always sucks to hold back a paper from submission, even when you know it is the right call. Super proud of my students who made this hard decision!.
5
3
135
I will be teaching Virtualization (for undergrads) in the fall. We cover basics of virtual machines, containers, and serverless technologies. Schedule: If you would like to give a guest lecture in the class, please email (vijayc@utexas.edu) or DM!.
7
17
129
Some folks are arguing that conferences *must* include alcohol because otherwise people will just go outside and get a drink. My model of conferences attendees is that they will go where their friends are, not where the alcohol is.
16
4
118
Something I wish was taught as a course: how to quickly read, understand, and contribute to a large code base. I would attend this course, and advise all my students to attend! . Unfortunately, not sure if this can be taught; maybe it only comes from experience?.
18
25
124
From Prof Jeff Naughton when I was a grad student @WisconsinCS: . "Life is like a DB query optimization problem. The important thing is to avoid the worst plans, even if you don't end up picking the optimal plan".
What’s the single best life advice you’ve ever been given?. Not a list. Tell me the best ONE. You have to choose.
3
16
128
Almost every CS research paper I see shared on social media is an arxiv paper. Why do we need publishers in CS again? Arxiv is free both to authors and the folks who read our papers. It is maintained well, and rarely down. We should redirect publisher $$$ to arxiv.
5
22
123
Happy to share our group @UTSASLab had two papers accepted to #SOSP2019! . Congratulations to @RohanKadekodi, @SekwonL, and @jayashree2912! . Both papers are about persistent memory, in collaboration with Taesoo Kim @gtcomputing. A quick thread on the main ideas in the papers:.
4
3
124
Fellow profs who try to give "fuses" to grad students trying to decide between offers: this is a bad tactic, and it spoils your name. A PhD advisor should look out for their student, and if they get an offer that would be better for them, you should encourage them to take it.
8
6
125
Excited and proud to share that my student @SekwonL was selected as one of the 2021 Microsoft Research PhD Fellows! Sekwon started in my group @UTSASLab in 2018 and is working on building storage systems for persistent memory.
We're excited to announce the 2021 Microsoft Research PhD Fellowship recipients! For over two decades, our PhD Fellowship has supported nearly 300 outstanding students pursuing exceptional research. Learn more about our 10 new fellows:
9
4
118
Wouldn't it be nice if the expectation for a systems PhD was:.1) build one big, novel system, make it practically usable by third parties.2) continue to maintain the system, give talks and educate people about it .3) write about their experiences updating and maintaining it.
Recording of the Noria stream is up! I think it went well — people seemed to follow the system design and the bug we ended up partially fixing, even though it was quite technical! If you're stuck at home and craving some research-y @rustlang, give it a try
7
6
120
Neurips might have been a research conference at one point, but it is becoming something else entirely. Like a carnival/festival for people interested in AI/ML? Coachella for AI?. This is all great, but I feel like we should maybe separate the research part into a diff conf.
3
5
118
I was just giving my students and postdocs advice about networking at #osdi2022 next week, and figured I'll tweet it. It will also go into the CS Asst Prof Handbook (!.
1
23
119
Excited to announce our paper, “Finding Crash-Consistency Bugs with Bounded Black-Box Testing”, was accepted to #osdi18 :) . First OSDI paper for @jayashree2912, @Ponnapalli95, @pandian4tweet, Ashlie Martinez, and @UTSASLab :).
10
4
117
All the systems researchers who consistently publish in the top conferences are also excellent writers (as in, they write beautiful prose). This is more than correlation: a paper that is easier to read and understand has a higher chance of getting accepted.
6
11
116
I've been seeing some people post that "we should be tolerant of those with other views". This depends on what the views are: if its vim vs emacs, sure. But if the view is that a group of people should have no rights, etc, we should *not* tolerate this.
4
15
110
Folks heading to #atc18 check out our work on TxFS, the Texas Transactional File System!. We built TxFS in only 5K LOC by utilizing the file-system journal to get atomic and durable transactions.
9
6
113
One of the pieces I read in the recent past from @zeynep that had a big impact on me. Basically, she says that a good society/system shouldn't need heroes: incentives should be aligned with expected good behavior. I've started looking at everything through that lens since.
0
42
101
Given that a lot of grad students will be doing visit days and trying to pick a school soon, I thought I'd retweet my old blog post about this and summarize the advice: trust your gut, avoid abusers, optimize for being happy.
3
23
118
A common way to discriminate against folks is to have a really strict set of rules, but be selective about who it is applied to. Then they can say "but they broke the rules"! Others from the in-power group breaking same rules get a pass.
Apparently @mmitchell_ai had violated their "codes of conduct.". I'd like to see which parts. Harassment and bullying seems to be condoned by the code of conduct because they sure seem to know how to protect those people and seems impossible to fire them.
1
25
107
He argued that weak papers would be ignored over time anyway, thus there is not a lot of cost to accepting the paper. On the other hand, you set back the field quite a bit with a false negative: a paper that should have been accepted, but instead rejected.
2
10
110
I think folks are imagining that Biden will win in November, wave some magic wand that makes all of this okay. This feels a little optimistic: Even if he wins, he has a ton of things on his plate, and I'm not sure where international students are in his priority list.
4
4
111
Systems papers a decade ago: We show how to build a reliable file system. (immediately gets into details). Systems papers today:.- this is why a file system is important.- this is why reliability is important.- now that we have established this is an important problem. .
9
9
109
Quick rant: reviewing the work .vs.reviewing the paper. I think what we should be doing in peer review is reviewing the *work*: the underlying ideas, whether they would have impact, whether folks would benefit from hearing of the work.
5
12
111
If you are working in distributed systems, I think one of the most impactful things you can work on right now to improve the life of lots of software engineers is working on ways to reduce pageable incidents when engineers are on call. For example, better monitoring.
3
24
111
From PC invitations, it seems like folks want to go back to in-person PC once the pandemic is over. Can we please not do this? . We have been doing online PCs perfectly well during the pandemic. Travel for a PC is fantastically wasteful. Let us network some other way.
13
7
106
I wish all conferences would release the list of Best Paper Nominees: papers which were considered for the best paper award. Often as good as the award papers, and easy way to recognize students without doing more work!.
5
9
109
When I say success in grad school is dependent more on attitude and character than intelligence, I mean stuff like this. Someone willing to ask "dumb" questions will progress faster than someone who doesn't. Not caring what others think of you is a superpower in some cases.
1
8
107
Didnt fsync to increase performance.
give me a horror story from your specialty in five words or less.
8
17
106
This diwali all my international grad students have gone home and are safely celebrating with their families :) Very thankful for that! Happy Diwali :).
1
3
111
Just had a paper rejected with .1) no updates in any reviews after the rebuttal.2) no post-rebuttal author-visible summary . Sure, it might have resulted in a ton of discussion among reviewers. Or not. I think you can see why I'm not hot on rebuttals :).
11
1
103
Imagine if theory folks said “We are going to solve p = np, but it’s going to take billions of dollars, so we are going to start a capped-profit company to do that” Add in statements about saving the world, solving the hardest challenges etc etc. They could call it OpenTheory.
3
22
107
Academics: .Number of citations is a terrible metric and we shouldn't use them . Also academics:.Look at how well my group/my university/my conf does so well on these metrics!.
8
6
100
I think systems sorely needs an open-access, free-for-authors, free-for-readers high-quality journal. Here's my proposal for one based on the PVLDB model and @cfiesler's comment that law journals like Harvard Law Review are run by students.
5
31
104
What used to be the Systems Slack is now the CS Research and Practice Slack! New name to reflect the broadening of scope, and to make everyone feel welcome :) . 1800 members and counting! Public and private channels on a variety of CS topics. Invite link:
1
24
103