Signal and Power Integrity Engineer, Graphcore
2+ days a week in office
Advanced AI systems and services
Be an early applicant
Advanced AI systems and services
201-500 employees
Be an early applicant
2+ days a week in office
201-500 employees
Company mission
To make machine learning faster, easier and more intelligent.
Role
Who you are
- You will be an expert user of high-speed PCB simulation tools and be familiar with industry best practice in high-speed digital system design
- 5+ years’ experience in high-speed digital circuit development, with a deep understanding of signal and power supply integrity
- Knowledge of high-speed circuit design requirements to ensure signal integrity and robust rejection of interference and crosstalk is essential
- Experience of high frequency (tens of GHz) SERDES design and validation
- Experience with ANSYS SIWave, HFSS, and/or Mentor Hyperlynx Signal and Power Integrity tools
- High speed processor system design with experience in interfaces such as SERDES, PCI, DDR, Ethernet, etc
- Design of BGA & connector pad launches, via structures and transmission lines for GHz signals
- This position will require extensive practical work in the lab, bringing up, debugging and validating designs
- Expertise in the use of vector network analysers, high speed digital oscilloscopes, signal generators
- A rigorous approach is required to make sure designs will work correctly the first time
- Independent working and the taking of responsibility, with excellent cooperative and communications skills to work in a small team environment
- Experience working in a small team, which requires a wide range of skills, an eagerness to participate in all parts of the development cycle from concept to testing in the lab and supporting manufacture and customers
What the job involves
- As a hardware signal integrity engineer you will develop, extract and simulate PCB structures to ensure that the high speed SERDES (112Gbps) and DDR interfaces perform correctly within the system budget
- You will work closely with the PCB and package substrate CAD engineers to design component signal launches from BGA components and connectors, layer transition via structures, and cable interconnect
- You will define the low-loss materials to be used for future designs and work closely with Graphcore’s system architecture team and suppliers to ensure the best possible performance is achieved
- Graphcore’s Hardware Engineering Team creates both products for sale in volume production and hardware test & validation systems based on Graphcore’s IPU processor
- These systems allow Graphcore’s customers to develop leading-edge machine-leaning solutions for training and inference
- Signal and Power Integrity specialist for the development of production systems and test boards for Graphcore’s family of processors
- Development of printed circuit board structures for high frequency (112Gbps, 16Gbps) SERDES interfaces, DDR interfaces
- PCB stackup definition
- System, PCB and substrate level simulation and analysis of signal and power integrity
- Graphcore Integrated Circuit device interface requirements definition, tool & vendor relationships
- PCB design requirement and constraint definition, PCB layout supervision
- Prototype bring-up, evaluation, validation & system test
Our take
Traditional hardware, mainly GPUs, used for training these models, comes at a high cost, both financially and environmentally, producing millions of dollars in expenses and hundreds of tons of greenhouse gases. In response, Graphcore is developing accelerator chips designed to expedite AI system training while significantly reducing power consumption.
AI accelerators like those crafted by Graphcore specialise in enhancing specific AI applications, such as artificial neural networks, deep learning, and machine learning. Utilising multicore designs and prioritising low-precision arithmetic or in-memory computing, these accelerators elevate the performance of large AI algorithms, achieving breakthrough results in natural language processing and computer vision.
Graphcore has ambitious plans to creat an AI supercomputer set to debut in 2024, aiming to be one of the most powerful in its class. The company is actively seeking strategic partnerships to bring this vision to fruition, signaling its commitment to shaping the future of AI computing.
Freddie
Company Specialist at Welcome to the Jungle
Insights
Company
Funding (last 2 of 6 rounds)
Dec 2020
$222m
SERIES E
Feb 2020
$150m
SERIES D
Total funding: $682m
Company benefits
- Work-Life Balance: We believe you should have a good work-life balance. Everyone's situation is different, so we offer flexible working arrangements and start and finish times, allowing you to choose when to work from home or come into the office
- Wellbeing: We want to ensure that our employees have the resources and support they need to prioritise their health and wellbeing
- Healthcare: We offer Private Medical Insurance (PMI) to our employees to help them take care of their health and that of their loved ones
- Progression: We prioritise learning and development as an integral part of our culture at Graphcore. Our Talent Development team is dedicated to equipping all our employees with the necessary tools, resources, and support to enhance their skills and reach their full potential in their careers
- Pension: We believe that our employees deserve the best when it comes to planning for their future
- Income Protection: Graphcore understands that unforeseen circumstances can happen, and we want to ensure our employees have the necessary financial security in place
Company HQ
Old City, Bristol, UK
Leadership
Nigel Toon
(CEO)Previously the President and CEO at picoChip Designs (processors for telecoms). As well as the VP and GM of Europe at Altera (semiconductors, merged with Intel).
Simon Knowles
(CTO)Founded two successful semiconductor companies, Element14 and Icera, for a combined value of over $1 billion.
Share this job
View 31 more jobs at Graphcore