Hello there, I’m Ben Cheng!😊

šŸ‘Øā€šŸŽ“ I’m a third year undergraduate student at the University of Toronto and am currently interning in Ericsson developing 5G ML application.

šŸ’• My research interests are in Machine Learning, High Performance Computing and Infrastructure Design. šŸ¤– I’m currently working as a research assistant at the Vector Institute / UHN and participating in a medical time-series machine learning project supervised by Professor Michael Brudno and Professor Rahul Krishnan. I’ve also participated in Energy System + Graph Theory research with Professor Yize Chen and crafted a publication studying Alberta Power Distribution.

Personal Experience

Ericsson 5G ML Development

–Implemented Deep Q-Network (DQN) to 5G SRS Scheduling and RL Algorithm to 5G Antenna Calibration, catering to network providers including Bell and CMCC

–Collaborated on the implementation of Embedded Real-time Communication C++ Application, establishing synchronization between Inference Model, RL Training Server and InfluxDB Time-Series Database

–Participated in the research, implementation and Field Test Time-Series analysis of AiXellerator application, leading to 20% boost in real-time downlink throughput and 10% boost in spectral efficiency

Energy System + Graph Theory Research

–Investigated and implemented a carbon emission tracing algorithm on Alberta’s electricity network, aiming to visualize and analyze provincial-level or national-level carbon emission distribution

–Utilized the Power Dispatch Model with BFS and LP algorithms to formulate, derive, and evaluate the contribution of individual generators and visualize the power distribution and consumption distribution

–Supervised by Professor Yize Chen at the University of Alberta

Time-Series Machine Learning Research

–Investigated the influence of Recurrent and Feedforward Neural Networks in predicting Tacrolimus Dose Requirement for liver-transplant patients, aiming to optimize the clinical medicine dose selection

–Conducted neural network application experiments on liver-transplant datasets from University Health Network

–Supervised by Processor Michael Brudno and Professor Rahul Krishnan at the Vector Institute

Wifi-Positioning + LLM Research

–Developed a spatial-aware cartography application supporting indoor position estimation, relative rooms position reasoning and auto-decision during measurement and training process

–Researched the potential of utilizing CSP algorithm for relative relationship deduction and LLM for description organization and decision making

–Supervised by Professor Khai Truong at the Dynamic Graphics Project lab

Centre for Social Services Engineering Full Stack Development

–Implemented a responsive search engine with GraphDB and React, featuring multi-condition filtering and pagination, and constructed a Material-designed user interface constructed by React and TypeScript

–Optimized and streamlined SPARQL queries with Lucene integration, enhancing system efficiency and reducing search time by more than 60%, with improved system robustness and scalability

–Collaborated on a full-stack project, connecting Ontotext Graph Database with NodeJS backend codebase, constructed efficient data pipelines for information sharing, with comprehensive test coverage using Jest library

Middleware Systems Research Group Software Development

–Designed and developed a modern web application using Django CMS and Nginx, featuring a custom-built search engine crafted with scikit-learn

–Implemented a GitHub Actions CI/CD pipeline for automating code integration, Docker container deployment on the testing server, and the periodic management of secure HTTP certificates

–Integrated the Google Scholar API with MySQL and Redis databases to update and manage the publication information of research projects and collaborators

ActEarn Startup Software Engineer

–Contributed to the development of an AI-integrated social application designed to foster lasting connections between students and alumni, utilizing GAN and LLM model to enhance user engagement and network growth

–Leveraged React Native to develop a responsive mobile chat interface, integrating backend services with SendBird for real-time messaging and Clerk for user authentication, enhancing user experience and operational efficiency