Responsibilities
Writing reusable, testable, and efficient production code in Machine and Deep Learning pipelines for
Computer Vision
Design and implementation of low-latency, high-availability, and performant applications
Assess and prioritize feature requests
Coordinate with internal teams to understand user requirements and provide technical solutions
Integration of user-facing elements developed by front-end developers with server-side logic
Integration of data storage solutions like databases, key-value stores, blob stores, etc.
Simultaneous, multiple technical project handling.
Experience- 4-7 years
Location: Technopark, Thiruvanathapuram
Work Mode: Full Time and work from Office
Skills And Qualifications
• Any one of the following combinations:
• (Python and C++) OR (Python and Rust) OR (Python and Java) OR (Python and Kotlin):
(should have developed computer vision models with any of Jax, Tensorflow 2 or
PyTorch. Both programming languages are must, and Python with C++ or Python with
Rust streams will be ranked higher.)
• (Good to have): Deployed models and Computer Vision data processing models and
pipelines with ONNX Runtime, TensorRT and DeepStream Runtime
• CUDA-C/OpenCL-C – big plus!
• Demonstrable Image Processing and Computer Vision problem-solving skills, understanding of
modern machine learning methods and deep learning architectures and paradigms, e.g., object
detection and segmentation, Image classification with various Deep Neural Networks,
conceptual understanding of zero-shot architectures in various computer vision tasks.
Candidates may be asked to code specific DNN architectures and various computer vision and
image processing routines during the interview process.
• Proficient with Linux Command Line (Must, since most of the work is on remotely connected edge AI/IoT devices)
• Understanding of fundamental design principles behind a scalable application• Understands and is able to write and consume or learn writing HTTP REST and Streaming APIs
using Python, java or Rust stacks.
• Understanding of the differences between multiple delivery platforms, such as mobile vs desktop, and optimizing output to match the specific platform