The Future of Deep Learning is Sparse.The Future of Deep Learning is Sparse.

The Future of Deep Learning is Sparse.The Future of Deep Learning is Sparse.


It all started in Cambridge, Massachusetts.

While mapping the neural connections in the brain at MIT, Neural Magic’s founders Nir Shavit and Alexander Matveev were frustrated with the many limitations imposed by GPUs. Along the way, they stopped to ask themselves a simple question: why is a GPU, or any specialized hardware, required for deep learning?

They knew there had to be a better way. After all, the human brain addresses the computational needs of neural networks by extensively using sparsity to reduce them instead of adding FLOPS to match them.

Based on this observation and years of multicore computing experience, they created novel technologies that sparsify and quantize deep learning networks and allow them to run on commodity CPUs – at GPU speeds and better. Data scientists no longer have to compromise on model design and input size, or deal with scarce and costly GPU resources. Their ground-breaking discovery became the foundation of Neural Magic.

The Company’s Vision
Efficient machine learning via sparsity. Neural Magic is making deep learning faster, affordable, and environmentally friendly with automated model sparsification technologies that minimize footprint and allow models to run on CPUs at GPU speeds. The future of deep learning is sparse. At Neural Magic, we make sparsity work.
 

Neural Magic Team


Brian Stevens Executive Chairman

Former CTO of Red Hat and Google Cloud, disruptor, biker.
Nir Shavit Co-Founder

MIT professor, innovator, tennis player.
Alexander Matveev CTO & Co-Founder

Former MIT research scientist, specialized in multicore algorithms and systems for AI.
John O'Hara Vice President Engineering

Engineering and operations lead, skier, mentor to engineers and peers.
Gaurav Rao Vice President Product

Former IBM Trusted AI exec, physics enthusiast, sports eccentric, dog lover and traveler.
Jeannie Finks Head of Customer Success

Model optimization, deep learning economics, yoga, killer sangria, adventure travel.
Saša Zelenović Head of Marketing

Go-to-market expert, helping data scientists experience the power of Neural Magic. Weekend beekeeper.
Mark Kurtz Machine Learning Manager

Research and engineering, model optimizations, volleyball, woodworking
Dan Alistarh Research Lead

Algorithms, math, tennis---not necessarily in that order.
Bill Nell Software Engineering Manager

JITs, compilers, program analysis, HPC.
Tyler Smith Software Engineering Manager

UT Austin PHD, ETH Zurich postdoc. Specializing in linear algebra computations & theoretical bounds on data movement.
Dan Huang Software Engineering Manager

A constructor (DevOps) and a breaker (QA). Passion for automation, quality, cooking and mushroom hunting.
Andy Linfoot Principal Software Engineer

JITs, Numerics, and HPC alchemy.
Tuan Nguyen Senior Machine Learning Engineer

Building software and deep learning models for fun and profit.
Karl Meissner Senior Software Engineer

Passionate builder of video game AI, compilers, vision system, financial real time data feeds, flight simulators and other machine learning systems. Likes cooking great meals, too.
Michael Goin Software Engineer

Data-driven optimization, AI evangelist, modular synthesis, baking.
Will Leiserson Software Engineer

MIT Ph.D. in making software run fast on many cores. Kerbal Space Program enthusiast. Dank meme connoisseur.
Sage Moore Software Engineer

Operating systems, scheduling, high performance computing, biking, powerlifting.
Kierstin Darragh Account Executive

Computer vision at the edge, cooking, true crime podcasts.
Benjamin Fineran Machine Learning Engineer

Model sparsifier, problem solver, sports enthusiast.
Kevin Escobar Rodriguez Full Stack Engineer

Full time full stack engineer, part-time snacker. Avid fan of games, animation, and comics.
Govind Ramnarayan Software Engineer

MIT PhD in Theoretical Computer Science, with publications in discrete algorithms, statistics, and complexity theory. In rehabilitation to become a practitioner.
Kyle Singer Software Engineer Intern

WUSTL PhD candidate specializing in runtimes that make it easier to write parallel code. Avid reader. Amateur coffee roaster.
Daniel Campos Engineering Intern

NLP aficionado, PhD Candidate UIUC, kiteboarder, winemaker.

Investors


comcast ventures
pillar
nea
andreesen horowitz
amdocs

Join Our Team


We are looking for talented and ambitious team members to help us shatter the hardware barriers holding back the field of machine learning. Are you ready to challenge the norms?