AI Systems Engineer (Edge Computing) (IRC216136)
Core Technical Skills
- Machine Learning and Deep Learning: Solid understanding of ML and deep learning principles, algorithms, and architectures. Familiarity with a range of models, including but not limited to CNNs, RNNs, and transformers, across various applications like image processing, natural language processing, and predictive analytics.
- Model Conversion and Optimization: Expertise in converting ML models to ONNX format for interoperability and further optimizing them using NVIDIA TensorRT for high-performance inference on NVIDIA Jetson devices.
NVIDIA Jetson Platforms: Hands-on experience with NVIDIA Jetson Nano, TX2, Xavier NX, or AGX Xavier, including setup, configuration, and performance tuning. Knowledge of Jetson software stack, including JetPack SDK.
- Dependency Management: Proficiency in solving dependencies and compatibility issues during model conversion and deployment, including library versions, CUDA, and cuDNN dependencies.
- Programming Languages: Strong programming skills in C++, with experience in Python being a plus. Proficiency in using PyTorch, TensorFlow, or other deep learning frameworks for model development and training.
- Environment and Tools: Familiarity with Docker and virtual environments for managing project dependencies. Experience with version control systems, preferably Git, for code collaboration and versioning.
Would be a plus
- Edge AI Applications: Understanding of Edge AI concepts and experience in deploying AI solutions on edge devices, optimizing for low power consumption and latency.
- Data Preprocessing: Skills in data augmentation, preprocessing, and transformation techniques to prepare datasets for efficient training and inference on edge devices.
- Benchmarking: Experience in conducting benchmarks to measure model performance (e.g., inference speed, accuracy, power consumption) on NVIDIA Jetson platforms.
- Troubleshooting: Ability to troubleshoot and debug hardware and software issues, including performance bottlenecks and memory constraints on edge devices.
Job Responsibilities
- Solution Porting:
Leading the efforts to port existing machine learning solutions to NVIDIA Jetson hardware, ensuring compatibility and stability.
Collaborating with ML engineers to understand the intricacies of the models and algorithms being ported.
- Performance Optimization:
Optimization of machine learning models and algorithms to leverage the unique hardware features of NVIDIA Jetson, such as GPU acceleration, to enhance performance and efficiency.
Implementation optimizations such as model quantization, pruning, and TensorRT conversion to improve inference speed and reduce model size.
- Integration and Testing:
Integration of optimized models into the target hardware environment, ensuring seamless operation within the customer's existing ecosystem.
Conduct thorough testing and validation to identify and resolve any performance bottlenecks or compatibility issues.
- Collaboration and Communication:
Serve as the primary liaison between the machine learning engineering team and hardware specialists, facilitating clear communication and collaboration across these disciplines.
Department/Project Description
Continuation of the project related to self-driving cars.
- Machine Learning and Deep Learning: Solid understanding of ML and deep learning principles, algorithms, and architectures. Familiarity with a range of models, including but not limited to CNNs, RNNs, and transformers, across various applications like image processing, natural language processing, and predictive analytics.
- Model Conversion and Optimization: Expertise in converting ML models to ONNX format for interoperability and further optimizing them using NVIDIA TensorRT for high-performance inference on NVIDIA Jetson devices.
NVIDIA Jetson Platforms: Hands-on experience with NVIDIA Jetson Nano, TX2, Xavier NX, or AGX Xavier, including setup, configuration, and performance tuning. Knowledge of Jetson software stack, including JetPack SDK.
- Dependency Management: Proficiency in solving dependencies and compatibility issues during model conversion and deployment, including library versions, CUDA, and cuDNN dependencies.
- Programming Languages: Strong programming skills in C++, with experience in Python being a plus. Proficiency in using PyTorch, TensorFlow, or other deep learning frameworks for model development and training.
- Environment and Tools: Familiarity with Docker and virtual environments for managing project dependencies. Experience with version control systems, preferably Git, for code collaboration and versioning.
Would be a plus
- Edge AI Applications: Understanding of Edge AI concepts and experience in deploying AI solutions on edge devices, optimizing for low power consumption and latency.
- Data Preprocessing: Skills in data augmentation, preprocessing, and transformation techniques to prepare datasets for efficient training and inference on edge devices.
- Benchmarking: Experience in conducting benchmarks to measure model performance (e.g., inference speed, accuracy, power consumption) on NVIDIA Jetson platforms.
- Troubleshooting: Ability to troubleshoot and debug hardware and software issues, including performance bottlenecks and memory constraints on edge devices.
Job Responsibilities
- Solution Porting:
Leading the efforts to port existing machine learning solutions to NVIDIA Jetson hardware, ensuring compatibility and stability.
Collaborating with ML engineers to understand the intricacies of the models and algorithms being ported.
- Performance Optimization:
Optimization of machine learning models and algorithms to leverage the unique hardware features of NVIDIA Jetson, such as GPU acceleration, to enhance performance and efficiency.
Implementation optimizations such as model quantization, pruning, and TensorRT conversion to improve inference speed and reduce model size.
- Integration and Testing:
Integration of optimized models into the target hardware environment, ensuring seamless operation within the customer's existing ecosystem.
Conduct thorough testing and validation to identify and resolve any performance bottlenecks or compatibility issues.
- Collaboration and Communication:
Serve as the primary liaison between the machine learning engineering team and hardware specialists, facilitating clear communication and collaboration across these disciplines.
Department/Project Description
Continuation of the project related to self-driving cars.
About GlobalLogic
GlobalLogic, a Hitachi Group Company, is a leader in digital engineering. We put people first. As part of our team, you will grow, be challenged, and expand your skill set working alongside highly experienced and talented people.In Ukraine, GlobalLogic is:
- one of the TOP-3 largest IT companies
- 6,000+ professionals
- 90%+ of our projects involve complex R&D
- fully autonomous offices are located in Kyiv, Kharkiv, Lviv, and Mykolaiv, along with 10 temporary mini-offices across Ukraine
What is GlobalLogic in numbers:
- 29,000+ engineers
- 20+ countries
- 500+ active clients
- 50+ product engineering centers
- Headquartered in Silicon Valley
Company website:
https://bit.ly/GlobalLogic-Ukraine
DOU company page:
https://jobs.dou.ua/companies/globallogic/
Job posted on
23 April 2024
45 views 1 application
45 views 1 application
Average salary range of similar jobs in
analytics →
Similar jobs
Senior Azure AI tools Developer at Kozak Agency
Ukraine
All jobs ML AI Kyiv All jobs GlobalLogic