Edward Schamp

Experience

• 

  China Young Entrepreneurs Association

  Nov 2015 - Sept 2018

  Founder

  - Started this non-profit to promote business and entrepreneurship to students in China. - Organised annual competitions for 3 years, gaining leadership and management skills.- Winning teams received prize money to kickstart their business ventures.- Upon my graduation CYEA was active in Beijing, Guiyang, Shandong and Shenzhen.

• 

  3DPTek

  Jun 2018 - Jun 2018

  Summer Intern

  - In a team of 3, focused on prototyping innovative components to maximise profitability.- Included an intense introduction to 3D printing, succeeded by learning its economic viability.- Equipment operated with a value of $5 million.

• 

  Joint research lab - Tianjin University and Procter and Gamble

  Jul 2018 - Aug 2018

  Summer Intern

  - Provided experimental assistance to develop an innovative product for the hair and body. - Operated lab equipment including the rheometer and industrial NMR machine.- Performed distillation experiments to measure detergent properties.

• 

  Southeastern Railway

  Jul 2020 - Aug 2020

  Web Development Intern

  - Designed a web module to automate various tasks using QR code scanning technology.- Worked to allow database compatibility for full integration in internal system.- Produced full report on the project, including financial benefits and further potential.- Coordinated directly with supervisor through online meetings and when necessary.

• 

  Imagination Technologies

  Apr 2021 - Sept 2021

  AI Research Placement

  - Optimised neural networks for specialised hardware (attention-based models).- Engineered and optimised 4 bit transformer models, obtaining promising results.- Communicated thoroughly by daily progress calls and weekly team meetings.

• 

  Imagination Technologies

  Oct 2022 - Feb 2024

  AI Researcher

  PyTorch projects:- 10x performance improvement for graph neural networks on the Neural Network Acceleratorthrough optimising the matmul and convolution operations performed in the network- Layer optimisation of language and vision transformers (Llama, Llama2, ViT) for the NNA- Main author of two filed patents to optimise sparse Graph Neural Network training andinference by partitioning the network layers, taking advantage of the sparsityHardware level projects:- Coded OpenCL kernels with 95% GPU utilisation for “gather” operations- Optimised the kernel execution and kernel fusing for heterogeneous networks running on theNNA and GPU, including diffusion networks, and transformer models for vision and textNetwork Deployment projects:- Implemented and deployed an adaptation of CodeLlama, quantised, for internal company use- Developed a chatbot with access to company documents using embedding vector similarity Show less

• 

  Aspire Software

  Mar 2024 - Dec 2024

  AI Architect

  I developed end-to-end machine learning solutions for the different software businesses inside of Aspire.

• 

  Google

  Jan 2025 - now

  Software Engineer - Core ML, Applied ML