@ReshabhSharma01 Profile picture

Reshabh K Sharma

@ReshabhSharma01

Everything is a compiler! PhD student @uwcse Opinions are my own!

Similar User
Programming Languages Mentoring Workshop photo

@PLMentoring

Chandrakana Nandi photo

@ChandrakanaNaN

Kirshanthan Sundararajah photo

@Kirshanthan

Santosh Nagarakatte photo

@santoshgnag

Ben Delaware photo

@GhostofBendy

Aditya photo

@_hiraditya_

Pavel Panchekha photo

@pavpanchekha

Pedro Abreu photo

@p_droabreu0

Uday Bondhugula photo

@ukrb84

Guannan Wei photo

@GuannanWei

Sameeran Joshi photo

@sameeran_joshi

vollmerm photo

@ma_vollmer

Harry Goldstein photo

@hgoldstein95

Pratyush Das photo

@ReikDas

Aviral Goel photo

@TheAviralGoel

Reshabh K Sharma Reposted

Our latest work on defenses against microarchitectural side-channel attacks.

Sophisticated side-channel attacks present new security challenges for cloud providers. Learn how Microsoft is exploring defenses against emerging attacks with principled microarchitectural isolation: msft.it/6013W5fTf

Tweet Image 1


Reshabh K Sharma Reposted

I like this framing. If prompts are programs, then adversarial prompts are malware.

I just co-authored a blog post arguing that #LLM #prompts are programs and require new research from #ProgrammingLanguages and #SoftwareEngineering. We need to rethink our #SoftwareTools to work for prompts blog.sigplan.org/2024/10/22/pro…



Reshabh K Sharma Reposted

I just co-authored a blog post arguing that #LLM #prompts are programs and require new research from #ProgrammingLanguages and #SoftwareEngineering. We need to rethink our #SoftwareTools to work for prompts blog.sigplan.org/2024/10/22/pro…


Reshabh K Sharma Reposted

Thanks to Nvidia for helping us bring compiler practitioners together at the first IICT workshop in Bengaluru! #IICT compilertech.org

Step into the future of compiler tech with @nvidia as a proud 𝐆𝐨𝐥𝐝 𝐓𝐢𝐞𝐫 𝐬𝐩𝐨𝐧𝐬𝐨𝐫! Join us on the 28th & 29th of September at @DayanandaOf for the Innovation in Compiler Technology Workshop 2024. Visit here- compilertech.org #CompilerTech2024 #IISc

Tweet Image 1


Reshabh K Sharma Reposted

Can user-intent formalization (clarification through specifications starting with tests) help users generate correct code from informal intent? TSE paper explores this question with a user study and empirical evaluation with TiCoder ieeexplore.ieee.org/document/10606…


Reshabh K Sharma Reposted

Thrilled to see my collaborator, Peli de Halleux, describe our project GenAIScript, a programming language for easily leveraging hashtag#GenAI in you scripts. #generativeAI #AI #scripting linkedin.com/events/genaisc…


Reshabh K Sharma Reposted

Azure Research in Cambridge, UK, is hiring two postdoctoral researchers --- in Security and Systems and in AI Security and Privacy: lnkd.in/eFxPJsBj lnkd.in/ewvq627Y Come work with us on exciting and important problems!


Reshabh K Sharma Reposted

Hi all, just a friendly reminder on the submission deadline...

Only two more days left for submissions of presentations and tutorials! visit: compilertech.org for more info. #llvm #gcc #rustc #ProgramAnalysis #EmbeddedSystems #Performance #SoftwareVerification #cgo #pldi



Reshabh K Sharma Reposted

FYI: IICT(Innovations In Compiler Technology) workshop will be held in Bangalore this September28-29. It will be a two day workshop with four tracks. Please see the CFP for more details. Travel assistance available to students. compilertech.org #compilers


Reshabh K Sharma Reposted

SafeAR will be hosted in Bellevue, Washington on October 21st 🏔️. We look forward to your great submission 🙌

Excited to co-organize the SafeAR workshop @ismarconf! We’ll explore how to reshape today’s AR development practices to address security & privacy risks. sites.google.com/umich.edu/safe… Apply to participate by July 26th, 2024!

Tweet Image 1


Reshabh K Sharma Reposted

Here's where we are with SIGPLAN-M for @PLDI 2024. Excited to have new mentees and mentors! We would love more mentors; please do sign up using the registration form on our website! If you're worried about seniority, remember that graduate students make wonderful mentors!

Tweet Image 1

Reshabh K Sharma Reposted

The video recording is now online (youtu.be/6fVUMJmJ2rM)

LLVM Social Bangalore is back with another online meetup! Title: Fixing the see-saw: Enabling precise optimizations in JIT compilers, efficiently Date: 18th May 2024 (Saturday) Time: 3:00 PM onwards Speaker: Prof. Manas Thakur, IIT Bombay RSVP: meetup.com/bangalore-comp… #llvm



Reshabh K Sharma Reposted

We need your feedback on an llvm.org website redesign! Please fill out this survey: docs.google.com/forms/d/1xr1Cl…


Reshabh K Sharma Reposted

Had a great time speaking to the BuzzRobot community about GPU security! Thanks to @sopharicks for the invitation and looking forward to engaging more with the community!

It was a very insightful lecture by @Tyler_UCSC for the @buZZrobot community about the AI security risks associated with GPU local memory leaks and which companies' devices were affected by this kind of vulnerability. Watch the lecture to learn about new security issues of LLMs.…

Tweet Image 1


Reshabh K Sharma Reposted

Today's the day for PNW PLSE! Take a look at our schedule of talks here: pnwplse.org!


Reshabh K Sharma Reposted

AI and Security? OpenSSF has a working group that handles everything at the intersection of the two. See a spotlight article on OpenSSF's blog for the working group: openssf.org/blog/2024/04/2…


The new Llama Guard in Llama3 dramatically outperforms its predecessor by achieving an F1 score of ~70, compared to 0, on the SPML prompt injection dataset. It is important to note that it only focuses on jailbreak detection but remains adaptable to custom safety criteria.

In addition to Llama 3, we also released new Meta Llama trust & safety tools featuring Llama Guard 2, Code Shield and Cybersec Eval 2 — plus an updated Responsible Use Guide & Getting Started Guide, new recipes and more. You can find more of these updates on the Llama website ➡️…

Tweet Image 1


Loading...

Something went wrong.


Something went wrong.