Only AI Jobs


Sr. Software Engineer- AI/ML, AWS Neuron Distributed Training

ID: 5371

Type: Full-time

Category: Others

Company Name: Annapurna Labs (U.S.) Inc.

Location: USA, CA, Cupertino - Cupertino - United States

Salary: 193,300.00 - 261,500.00 USD annually

Visit company vacancy
Job Description

Annapurna Labs designs silicon and software that accelerates innovation. Our custom chips, accelerators, and software stacks enable us to tackle unprecedented technical challenges and deliver solutions that help customers change the world. AWS Neuron is the complete software stack powering AWS Trainium (Trn2/Trn3), our cloud scale Machine Learning accelerators and we are seeking a Senior Software Engineer to join our ML Distributed Training team.

In this role, you will be responsible for the development, enablement, and performance optimization of large scale ML model training across diverse model families. This includes massive scale pre-training and post-training of LLMs with Dense and Mixture-of-Experts architectures, Multimodal models that are transformer and diffusion based, and Reinforcement Learning workloads. You will work at the intersection of ML research and high performance systems, collaborating closely with chip architects, compiler engineers, runtime engineers and AWS solution architects to deliver cost-effective, performant machine learning solutions on AWS Trainium based systems.

Key job responsibilities
You will design, implement and optimize distributed training solutions for large scale ML models running on Trainium instances. A significant part of your work will involve extending and optimizing popular distributed training frameworks including FSDP (Fully-Sharded Data Parallel), torchtitan and Hugging Face libraries for the Neuron ecosystem.
A core focus of this role involves developing and optimizing mixed-precision and low-precision training techniques. You will work with BF16, FP8, and emerging numerical formats to maximize training throughput while maintaining model accuracy and convergence quality. This requires implementing precision aware training strategies, loss scaling techniques, and careful gradient management to ensure training stability across reduced precision formats. Understanding the tradeoffs between computational efficiency and numerical fidelity is essential to success in this position.
Beyond precision optimization, you will profile, analyze, and tune end-to-end training pipelines to achieve optimal performance on Trainium hardware. You will partner with hardware, compiler, and runtime teams to influence system design and unlock new capabilities. Additionally, you will work directly with AWS solution architects and customers to deploy and optimize training workloads at scale.


About the team
Annapurna Labs was a startup company acquired by AWS in 2015, and is now fully integrated. If AWS is an infrastructure company, then think Annapurna Labs as the infrastructure provider of AWS. Our org covers multiple disciplines including silicon engineering, hardware design and verification, software, and operations. AWS Nitro, ENA, EFA, Graviton and F1 EC2 Instances, AWS Neuron, Inferentia and Trainium ML Accelerators, and in storage with scalable NVMe, are some of the products we have delivered, over the last few years.

Basic Qualifications

- Bachelor's degree in computer science or equivalent
- 5+ years of non-internship professional software development experience
- 5+ years of programming with at least one software programming language experience
- 5+ years of leading design or architecture (design patterns, reliability and scaling) of new and existing systems experience
- 5+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
- Experience as a mentor, tech lead or leading an engineering team
- Experience in machine learning, large scale training with LLMs and expertise in Pytorch.

Preferred Qualifications

- Master's degree in computer science or equivalent
- Experience in computer architecture
- Previous software engineering expertise with Pytorch/Jax/Tensorflow, Distributed libraries and Frameworks, End-to-end Model Training.

Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

Los Angeles County applicants: Job duties for this position include: work safely and cooperatively with other employees, supervisors, and staff; adhere to standards of excellence despite stressful conditions; communicate effectively and respectfully with employees, supervisors, and staff to ensure exceptional customer service; and follow all federal, state, and local laws and Company policies. Criminal history may have a direct, adverse, and negative relationship with some of the material job duties of this position. These include the duties and responsibilities listed above, as well as the abilities to adhere to company policies, exercise sound judgment, effectively manage stress and work safely and respectfully with others, exhibit trustworthiness and professionalism, and safeguard business operations and the Company’s reputation. Pursuant to the Los Angeles County Fair Chance Ordinance, we will consider for employment qualified applicants with arrest and conviction records.

Our inclusive culture empowers Amazonians to deliver the best results for our customers. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, including support for the interview or onboarding process, please visit https://amazon.jobs/content/en/how-we-hire/accommodations for more information. If the country/region you’re applying in isn’t listed, please contact your Recruiting Partner.

The base salary range for this position is listed below. Your Amazon package will include sign-on payments and restricted stock units (RSUs). Final compensation will be determined based on factors including experience, qualifications, and location. Amazon also offers comprehensive benefits including health insurance (medical, dental, vision, prescription, Basic Life & AD&D insurance and option for Supplemental life plans, EAP, Mental Health Support, Medical Advice Line, Flexible Spending Accounts, Adoption and Surrogacy Reimbursement coverage), 401(k) matching, paid time off, and parental leave. Learn more about our benefits at https://amazon.jobs/en/benefits.



USA, CA, Cupertino - 193,300.00 - 261,500.00 USD annually

Company Information

Company Name: Annapurna Labs (U.S.) Inc.

Company Website: https://annapurna-labs.com

Company Address: 10201 Torre Avenue, Cupertino, California 95014, USA

Annapurna Labs (U.S.) Inc. is the U.S. corporate entity of Annapurna Labs, an engineering organization originally founded as a semiconductor and systems-on-chip (SoC) design company that was acquired by Amazon in 2015 and subsequently integrated into Amazon Web Services (AWS). The business is organized around the design and delivery of custom silicon, hardware acceleration engines, and tightly integrated hardware–software platforms used to improve performance, security, and efficiency of cloud infrastructure. Annapurna Labs operates as an engineering and product-development organization whose principal output is purpose-built silicon, firmware, and supporting software stacks that are deployed inside AWS data centers and products. Company overview: Annapurna Labs began as a privately held microelectronics and chip-design startup with an emphasis on low-power, high-density compute and specialized accelerators. After acquisition by Amazon, the group continued as an internal Amazon/AWS engineering organization, focused on developing devices and subsystems that address bottlenecks in cloud servers and networking equipment. The organization is widely credited in official Amazon/AWS materials and public communications with designing and delivering several generations of custom AWS silicon and the enabling hardware subsystems that underpin notable AWS services and instance families. Core business activities: Annapurna Labs’ core activities center on end-to-end silicon and platform engineering: architecture, RTL and logic design, physical implementation, verification, firmware, board design, and the software integration required to bring custom chips into production at hyperscale. Key technical specializations include SoC and CPU core design, network and storage offload engines, hardware virtualization and security enclaves, high-speed interconnects, and low-level system software to exploit hardware capabilities. Instead of operating as a customer-facing chip vendor, the organization’s outputs are integrated into AWS infrastructure and services—its primary “customers” are AWS teams and the cloud customers that rely on AWS services. Main products and services: Public and official AWS communications associate Annapurna Labs with multiple internal silicon and platform programs that have been announced or described by Amazon. Notable outcomes attributable to the group include the AWS Nitro System, a collection of dedicated hardware and firmware components that offload virtualization functions (such as network and storage I/O and security isolation) from host CPUs, enabling higher performance and stronger isolation for EC2 instances; and the Graviton family of Arm‑based processors (Graviton and subsequent Graviton2/Graviton3 generations), which provide high performance per watt and are used in a wide range of EC2 instance types. Annapurna Labs’ technology has also contributed to AWS’s custom accelerators and specialized chips for inference and other workloads, along with supporting board and subsystem designs used across AWS server fleets. Operational model and customers: Following its acquisition, Annapurna Labs functions primarily as an in-house R&D and product engineering organization for Amazon/AWS rather than as a standalone commercial vendor selling chips directly to third parties. Its deliverables are integrated into AWS compute, storage, and networking products to improve throughput, reduce cost and power consumption, and enable new service capabilities. Through these integrations, AWS customers—enterprises, startups, and public-sector users—indirectly benefit from Annapurna Labs’ designs in the form of new instance types, specialized acceleration options, and platform-level features (for example, enhanced virtualization security provided through Nitro). Technology and integration focus: Annapurna Labs emphasizes hardware–software co-design, meaning the group develops chips and the low-level firmware and drivers required to expose their capabilities to higher-level cloud services. This includes designing secure firmware stacks, silicon-based root-of-trust features, and hardware acceleration engines for packet processing, storage offload, and cryptographic operations. The designs are intended to scale in rack- and data-center-level deployments; consequently, the organization also contributes to manufacturing qualification, supply-chain integration, and operational tooling required to bring new silicon into production across AWS’s global infrastructure. Strategic impact and role inside AWS: The creation and deployment of custom silicon and Nitro-class offload engines have been described in AWS materials as foundational to certain performance, security, and cost-efficiency improvements across Amazon’s cloud offerings. Annapurna Labs’ designs have enabled AWS to introduce distinct product capabilities—such as high-density, energy-efficient Arm-based compute instances and dedicated hardware paths for I/O and isolation—that differentiate AWS services in performance-per-dollar metrics and security posture. While Annapurna Labs itself is not a public-facing product company, its engineering outputs are a strategic enabler for AWS’s infrastructure roadmap and service portfolio. Publicly disclosed collaborations and communications: Details about Annapurna Labs and its projects have been shared in Amazon press releases, AWS blog posts, technical presentations, and regulatory filings since the acquisition. Those official channels describe the group’s role in developing the Nitro System and custom processors, as well as the subsequent use of that technology in EC2 instance families and other AWS services. As an entity, Annapurna Labs (U.S.) Inc. is a legal corporate unit tied to Amazon’s organizational and intellectual-property structure, responsible for managing U.S.-based engineering operations, compliance, and aspects of the commercial integration of its technologies within AWS. Limitations on external sales: Unlike independent semiconductor companies that market chips to a broad set of third-party OEMs, Annapurna Labs’ outputs are primarily intended for integration into Amazon/AWS infrastructure. As a result, most technical and product announcements emphasize the benefits delivered via AWS services rather than standalone Annapurna-branded products sold directly to external customers.
Visit company vacancy