Associate Technical Architect - Platform Job in Quantiphi

Role: Platform Engineer

Work location: Mumbai/Bengaluru

Overview:
A Machine Learning/Deep Learning Engineer who has experience with scaling model training from one GPU to multiple GPUs. Must have familiarity with popular Deep Learning Frameworks and distributed training paradigms to help customers with optimally using their hardware.

Skillset Needed:

• Experience with Multi-GPU training(Dask ML)
• Familiarity with Tensorflow, PyTorch, other popular Deep Learning Frameworks with an experience in using and implementing distributed training strategies like DDP, libraries like Horovod
• Understanding of Data Parallelism and Model Parallelism
• Understanding of algorithmic and engineering challenges, benefits, bottlenecks and trade-offs when scaling to larger number of GPUs
• Proven experience with Containers (Docker, Kubernetes) Web Services (SOAP/REST) and Scalable Storage(HDFS/Ceph)
• Familiarity of building MLOps pipelines to support development, experimentation, continuous integration, continuous delivery, verification/validation, and monitoring of AI/ML models
• Good knowledge of CI/CD pipelines and tools (GitLab or similar)
• Strong hands-on experience coding with Python
• RAPIDS - https://rapids.ai/
• Power users of Numpy, scikit learn, Pandas
• Dask - https://tutorial.dask.org/00_overview.html
• ML and SQL knowledge
• Experience with Dataframes and Arrays on GPUs
• Experience in one or more of the following use cases, strong in at least one:
  • NLP: Transformers, NeMo, Jarvis
  • CV: Object detection, Image segmentation, IVA - TLT, DeepStream
  • ASR/TTS: NeMo
  • Recommender Systems Merlin