Strong programming using Python/Objective C/Swift.
Good exposure to Deep Learning libraries like PyTorch and TensorFlow.
Thorough experience with common NLP algorithms and applications, including tokenization, language modelling, text decoding, text classifier etc.,
Preferred Qualifications
Actively exercising machine learning techniques, including implementing basic optimizers and regularizations; formulating and implementing loss functions for given tasks, applying insightful parameter tuning in model training and evaluation, reproducing and revising state-of-the-art research experiments; addressing accuracy and performance bottlenecks in distributed training and inference, and benchmarking different model compressing methods
Actively programming with high-quality codes across complex and large repositories, including implementing new optimizer or loss functions, algorithm optimizations for critical performance or accuracy, refactoring or redesigning of existing code-base for improved robustness and flexibility, processing large data set efficiently and accurately, establishing automated synthesis/training/evaluation pipeline across multiple sets of internal and external APIs, and applying common object-oriented design patterns
Experience of multi-modal modeling, presenting plans, progress, and results or demos regularly and concisely
Experience in NLP modeling for Indian languages is desired.