ML Converter Developer
Spark your Imagination through the power of technology
We are Sony Semiconductor IL, formerly Altair Semiconductor. We believe in pioneering the future through the power of creativity. As 5G IoT expands to include all types of connected “things”, our ultra-low-power and ultra-small chipset solutions are turning Cellular 5G IoT into reality.
Sony Semiconductor IL is an integral part of Sony Semiconductor Solutions Corporation, creating a company culture that combines the Sony legacy, values and technology with our own vision and market leadership in the fields of Cellular 5G IoT and AI.
We proudly provide millions of chipsets to customers around the world connecting to the world’s top, tier-1 cellular networks.
Believe that anything you can imagine, you can make real.
Your journey starts with us:
ML Converter Developer
About the DNN group
Sony Semiconductor Israel’s DNN group (Deep Neural Network group) is developing a leading-in-its-segment deep learning inference processor for edge applications.
We have a fundamental understanding of how neural networks operate. Our edge device runs the networks with almost no loss in accuracy and minimal power consumption. With our team’s expertise in SW and HW architecture, to develop a product that has the potential to be a landmark in computer technology.
The DNN group is a mixture of experts from various fields such as ML researchers, Alg, SW, FW, and HW, with the highest standard of executing complex projects. We have a proven track record of bringing up multiple projects, working and collaborating with the world industry giants
About The Position
The DNN group is looking to hire a motivated and skilled SW developer. In this role, you’ll have a direct impact on our product. You’ll work on our ML converter that lies at the core of our technology. We take neural networks from industry-standard frameworks (PyTorch/Tensorflow), do our magic keeping the network mathematically equivalent, but make it fit into our edge device. The work involves algorithmic problem solving and advanced software design skills. The code is written mostly in Scala but we are transitioning to Kotlin. We invest much thought into automation and packaging for smoother integration and delivery.
An ideal candidate
Here are two examples of an ML enthusiast, passionate SW developer that might fit
The discoverer: you are an opinionated thinker, independent, and auto deductive. At first, you’ll take nothing for granted and aim to know the details of everything you do, but also broaden your comfort zone. Keeping a fine balance between BFS/DFS of your work and tasks is an art you master. Once you are comfortable in your position and feel confident you start exploring what can be done better. You’ll collaborate with your environment and strive to enrich code, methods, tools, and whatever you see fit for an upgrade.
The sentinel: you are a pillar of knowledge. You are known for your wise advice and great problem-solving skills. Your programming language is your weapon of choice and you are a master of it. You like challenges, you like to solve puzzles. You prefer to be handed a healthy well-defined feature over a micro detailed exhausting TODO list. You prefer to have the freedom to figure out how to break it into tasks, test it and integrate it by yourself. You are happy to defend your methods and often you do at the start, but as people get to know you, your esteemed reputation allows you more freedom.
These are just two examples of characters that might fit into our group. There are many more off course
- Design and develop a multidisciplinary system for machine learning
- Manipulate neural networks graphs, exploring them and performing rewrites on them
- Own projects and features, break tasks into sub-tasks.
- B.Sc in Computer Science, Electrical Engineering or equivalent
- 5+ years of R&D experience in Kotlin/Java/Scala
- Independent, hard-working, and committed
- Great interpersonal skills, a team player and honest
- Experience in Deep Learning and Deep Learning development frameworks
- Packaging with Docker/other containers