DNN Model optimization Expert Engineer

We build accelerated runtime for on-device execution of GenAI LLM/LVM large models. Our aim is to create world s best accelerated On-Device ML inference & training platform. Team focuses to identifying some of the best techniques to accelerate neural networks by utilizing the underlying SoC capabilities. Our technology is a fusion of both machine learning / deep learning & core system concepts. We have built the first version of neural acceleration platform in Samsung which is far ahead in terms of system KPIs in comparison with the other competitor platforms. We are now securing this area with key patents and working towards extending this to multiple others areas.

Following are the goals of Neural Acceleration team:
Design and Develop a machine learning on device runtime platform/framework that is ahead of any other platforms today with best system KPIs without compromising quality.
Make the platform re-usable and scalable by extending the same to cloud. Work Profile:

• Responsible for design and implementation of neural acceleration platform with support for accelerating the networks on one or more processing units.
• Lead, design & develop machine learning algorithms for use case realizations in mobile devices.
• Ensuring team is building high quality code via a predicable process.
• Along with product owners, Managers, and team leads to refine and prioritize the work backlog
• Identify opportunities to leverage the capabilities of existing platform to take it to the next level.

Necessary Skills / Attributes:

• Ph.D in machine learning or a related field; or advanced degree and equivalent industry experience.
• Experience in model optimization using compression/Quantization and Neural architecture search and also having good knowledge on LLM and LVM specifically for ondevice.
• Theoretical & practical knowledge of machine learning/ deep learning experience. Excellent track of research excellence, publications & patents.
• Proficiency in Python, Linux, C++, various shallow/deep learning framework (Pytorch/Tensorflow/AI-Edge)
• Proficiency in machine learning algorithms, statistical analysis, quantitative analysis and deep learning algorithms like convolutional neural network, recurrent neural network etc.
• Expertise in knowledge in writing multi-threading, multicore (Task parallelization, TPL, GPU) programs.
• Excellent knowledge in processor architectures (ARM (big.LITTLE.) (v7 & v8))
• Expertise in one of the server side machine learning frameworks.
• Prior experience on distributed computing, compiler development & understanding of low level optimizations is added advantage.