Marcelo Orenes-Vera

CS Graduate Student at Princeton University
Computer Architecture

[NVIDIA @nvidia]

Open models are becoming the foundation for modern AI research.

At #ICML2026, 145 accepted papers cite NVIDIA Nemotron models and datasets, 74 NVIDIA papers were accepted, and ~2,000 accepted papers cite NVIDIA GPUs.

[Marcelo Orenes-Vera @orenes95]

More compute alone isn’t enough, we have to use it more efficiently.

NVIDIA’s Bryan Catanzaro explains how Nemotron, open models and full-stack co-design are shaping the future of AI.

Worth a watch: https://t.co/wEyS9jpniA https://t.co/APSy9pgMKv

[Marcelo Orenes-Vera @orenes95]

@tcuarch Thanks to the organizers of the PhD Forum!

[IEEE TCuArch @tcuarch]

MICRO'23 PhD Forum (https://t.co/SC84fg9DOR) was successfully held on Monday!
Here's a talk by Marcelo Orenes Vera (PhD candidate at Princeton).
Marcelo is currently on the job market for academia.
Please contact him if your school is hiring!

https://t.co/pH6zcil4AC

[PULP Platform @pulp_platform]

Reporting from MICRO in Toronto, here is the talk on "AutoCC: Automatic Discovery of Covert Channels in Time-Shared Hardware" with Marcelo @orenes95 and our Nils @niwist during the poster session. Check out the repo https://t.co/G7fldVi3d2 and the paper: https://t.co/wRgtkHgFet

[Marcelo Orenes-Vera @orenes95]

@gsneha261 @Princeton @EPrinceton @euanashley @BenedictPaten Congrats! Welcome!

[Martonosi Research at Princeton @MartonosiGroup]

🚨🚨MRM group member @orenes95 won the Student Research Competition at @MicroArchConf this week for his work: Scalable Distributed-data Architecture for Memory-bound Applications 🚨🚨 - Congrats Marcelo!! 🎉🎉

[Marcelo Orenes-Vera @orenes95]

I'm excited to be presenting this Wednesday at ISCA'22 our work on "Designing and realizing scalable latency tolerance for manycore SoCs" (open access https://t.co/XOye2uWBmP)

Thanks to my collaborators @JuanLuisAragon @JBalkind @anindamanocha @DavidWentzlaff @margmartonosi

[Columbia SLD Group @ColumbiaSld]

The official program of OSCAR 2022, the new workshop on open-source hardware, is out!
Check it out at https://t.co/I8NrrZSgCw
#OpenSourceHW #ISCA2022

[Marcelo Orenes-Vera @orenes95]

I'm excited to be presenting at the 1st Open-Source Computer Architecture Research (OSCAR) workshop this saturday in NYC (co-located with ISCA)
Where are presenting AutoSVA (https://t.co/ashtwuuWAo) at the verification session. Full program at https://t.co/clFVbimx2M

[Marcelo Orenes-Vera @orenes95]

@artolamola Lo hay, está en crecimiento, pero tiene ya alojamientos en algunos sitios: https://t.co/9c5vqqg9FK 💻🌞

Puedes preguntar si tienes pensada alguna zona en particular a @ubiworkCo

[Marcelo Orenes-Vera @orenes95]

@carlosjonayss @LissetteIbnz @artolamola Lo hay, está en crecimiento, pero tiene ya alojamientos en algunos sitios: https://t.co/9c5vqqg9FK 💻🌞

[Marcelo Orenes-Vera @orenes95]

@cerebras @MicroArchConf Great talk today by Sean Lie, very inspiring about how much compound speedup one can get by fully co-designing. I really enjoyed watching it at #MICRO54

[YosysHQ @YosysHQ]

👨‍💻Check out AutoSVA, an interesting tool by Marcelo Vera that can automatically generate formal test benches. https://t.co/BYjTjv150j https://t.co/2ZmXn2rq8K

[Marcelo Orenes-Vera @orenes95]

@YosysHQ Thanks @YosysHQ for your collaboration in providing the powerful SBY solver to prove the AutoSVA generated properties!🚀

[Cerebras @cerebras]

Today we introduced the world’s first brain-scale #AI solution at @hotchipsorg 2021, enabling a single CS-2 to support 120 trillion parameter models. We continue to push the boundaries of what’s possible in AI & unlock extreme-scale model potential! https://t.co/cXrhVv0uIO

[Marcelo Orenes-Vera @orenes95]

@ubiworkCo Hard to decide one! I would say Gaucín in Málaga

[Cerebras @cerebras]

Look who’s on the cover of @IEEESpectrum’s July issue! Read more about supersize AI powered by our record breaking 2.6 transistor Wafer Scale Engine #AICompute #AIResearch #Innovation
https://t.co/vp1MWTVWQf

[Cerebras @cerebras]

State-of-the-art #deeplearning should be simple and accessible to every #ML practitioner. That’s why the CS-2 provides cluster-scale performance with the programming ease of a single machine. Read Deep Learning – Programming at Scale to learn more https://t.co/85HFFbCLb5

[Marcelo Orenes-Vera @orenes95]

@ubiworkCo All sorts of great ideas came from that brainstorming