Responsibilities: 

Model architecture optimizations : optimize latest LLM and GenAI model architectures for NPUs, which involves reimplementing basic building blocks of models for NPUs  

Model Training and Fine-tuning: Fine-tune pre-trained models on specific tasks or datasets to improve performance. Implement state-of-the-art LLM training techniques such as Reinforcement Learning from Human Feedback (RLHF), ZeRO (Zero Redundancy Optimizer), Speculative Sampling, and other speculative techniques. 

Data Management: Handle large datasets effectively. Ensure data quality and integrity. Implement data cleaning and preprocessing techniques. Hands-on with EDA is a plus. 

Model Evaluation: Evaluate model performance using appropriate metrics. Understand the trade-offs between different evaluation metrics.  

LLM metrics: Sound understanding of various LLM metrics like MMLU, Rouge, BLEU, Perplexity etc. 

AWQ: Understanding of Quantization is a plus. Knowledge on QAT will be a plus. 

Research and Development: Stay up to date with the latest research in NLP and LLMs. Implement state-of-the-art techniques and contribute to research efforts. 

Infrastructure development : For coming up with new optimization techniques to minimize ONNX memory footprint, export time optimizations. 

Collaboration: Work closely with other teams to understand requirements and implement solutions. 

Required Skills and Experience: 

Deep Learning Frameworks: Hands-on experience with PyTorch at a granular level. Familiarity with tensor operations, automatic differentiation, and GPU acceleration in PyTorch. 

NLP and LLMs: Strong understanding of Natural Language Processing (NLP) and experience working with LLMs. 

Programming: Proficiency in Python and experience with software development best practices. 

Data Handling: Experience working with large datasets. Familiarity with data version control tools is a plus. 

Education: A degree in Computer Science, Machine Learning, AI, or related field. Advanced degree is a plus. 

Communication: Excellent written and verbal communication skills. 

Work experience : Open, 4 – 10 years of relevant experience.