Systems Software Engineer (DeepSparse Engine)

---

Job description

Neural Magic is an early stage AI software company pioneering a new way for organizations to run deep learning applications without specialized hardware accelerators. Based on decades of parallel computing research at MIT, Neural Magic has developed a software platform for deep learning inference that allows developers to sparsify and quantize deep learning models to minimize footprint and run on CPUs at GPU speeds. Please look through our website [ and GitHub repos [ to get a feel of what we are about.

Founded by an award-winning team of computer scientists out of MIT, Neural Magic is a venture-backed company headquartered in Davis Square, Somerville, MA. Neural Magic investors include Amdocs, Andreessen Horowitz, Comcast Ventures, NEA, and Pillar VC.

If you are someone who wants to contribute to solving challenging technical problems at the forefront of deep learning, this is the role for you.

We are looking for engineers excited to work with machine learning, parallel and concurrent algorithms, interested in working closely with the founding team, and who want to contribute to solving challenging, high-performance computing problems.

RESPONSIBILITIES

* Write robust and modern C++, working throughout the Deep Sparse engine. Depending on your interests and skills, this may include working on high-performance machine learning primitives, just-in-time compilation, graph algorithms, performance analysis and modeling, numerical methods, and scheduling algorithms.
* Review code and help develop best practices for the team.
* Work closely with machine learning teams to attain high performance for neural networks in the engine.

Requirements

You must have one of the following:

* BS or MS in computer science, or related field and 2+ years of experience developing software
* PhD in computer science, computer engineering, applied mathematics, or related field

We would prefer if you have experience in at least some of the following:

* Modern C++
* Deep neural network models and techniques
* High performance computing
* Mathematical software, especially linear algebra and signal processing
* Low-level assembly and SIMD programming
* Compilers
* Multicore programming, parallelism and concurrency

COMMUNICATION & CULTURE

* A self-directed practitioner who learns fast
* Strong communications skills with both technical and non-technical team members
* Strong sense of project ownership and personal responsibility
* Genuine interest in learning
* Open to feedback from peers
* Willing to work outside of comfort zone

Neural Magic is an equal opportunity employer. All applicants will be considered for employment without attention to race, color, religion, sex, sexual orientation, gender identity, national origin, veteran or disability status.