Sr Software Development Engineer

- 4+ years of non-internship professional software development experience
- 4+ years of programming with at least one software programming language experience
- 4+ years of leading design or architecture (design patterns, reliability and scaling) of new and existing systems experience
- Experience as a mentor, tech lead or leading an engineering team
- Bachelor's degree in computer science or equivalent
- 4+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
- M.Sc. or PhD Degree in computer science, engineering, statistics, mathematics or related field
- Proficiency in the TensorFlow and/or PyTorch frameworks
- Strong working knowledge of C++ programming language;
- Strong working knowledge of Python programming language
- Experience in developing highly scalable, fault-tolerant, distributed systems
- Experience with multi-threaded asynchronous development in C++
- Solid understanding of machine learning techniques and concepts
- Experience with High Performance Computing systems
- Experience with CUDA programming
- Experience with Deep Learning models for Natural Language Understanding or Computer Vision tasks;
- Experience with Linux kernel system calls or POSIX API (process control, communication, and device management)

Amazon is committed to a diverse and inclusive workplace. Amazon is an equal opportunity employer and does not discriminate on the basis of race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status. For individuals with disabilities who would like to request an accommodation, please visit https://www.amazon.jobs/en/disability/us.

Our compensation reflects the cost of labor across several US geographic markets. The base pay for this position ranges from $151,300/year in our lowest geographic market up to $261,500/year in our highest geographic market. Pay is based on a number of factors including market location and may vary depending on job-related knowledge, skills, and experience. Amazon is a total compensation company. Dependent on the position offered, equity, sign-on payments, and other forms of compensation may be provided as part of a total compensation package, in addition to a full range of medical, financial, and/or other benefits. For more information, please visit https://www.aboutamazon.com/workplace/employee-benefits. This position will remain posted until filled. Applicants should apply via our internal or external career site.
AWS AI is looking for world-class software developers to join the Deep Learning cross-framework team. In this organization, you will be responsible for contributing extensions to the TensorFlow and PyTorch machine learning frameworks and for developing cross-framework solutions to support training of Deep Learning models at scale, involving thousands of accelerators. You will be working in a fast-paced, cross-disciplinary team of engineers and researchers who are leaders in the field. You will take on challenging problems, elicit requirements, and deliver innovative solutions into production that consolidate the AI team as thought leaders in the space.

Key job responsibilities
As a Software Development Engineer in the SageMaker Engines team, you will be responsible for:
- Developing innovative solutions for supporting Large Language Model training in a cluster of nodes;
- Implementing model parallelism methods such as pipeline and tensor parallelism as extensions to the PyTorch framework;
- Implementing sharding of the model training state, activation checkpointing/offloading and other memory saving techniques;
- Optimizing distributed training by profiling, identifying bottlenecks and addressing them by improving compute and network performance, as well as finding opportunities for better compute/communication overlap;
- Optimizing communication collectives for the AWS network infrastructure;


About the team
The SageMaker Engines team develops technology for supporting training of Deep Learning models at large scale. This entails implementation of model parallelism and memory saving techniques to allow training of models across accelerators as well as implementation of network communication collectives optimized for the AWS infrastructure.