Chengdong Liang
Process Engineer at Pharmaron
Dongli District, Tianjin, China
Summary
🤩
Rockstar🎓
Top SchoolChengdong Liang is a process engineer with 8 years of experience specializing in pharmaceutical CMC, technology transfer and scale-up at Pharmaron, where he bridges lab process development and commercial plant operations. He leads MPI/MBR design, equipment selection and qualification, performs risk assessments, and interprets complex thermal and calorimetry data (TGA-DSC, ARC, RC1) to troubleshoot deviations and size vents and heat removal systems. Proficient in mass and heat transfer calculations, CFD simulation, DOE and statistical trend analysis, he also develops custom automated data-processing tools that speed decision-making during scale-up. Chengdong has driven introduction of new technologies such as flow chemistry and PATs while acting as an internal audit SME for cGMP compliance. Uncommonly for a chemical engineer, he contributes to open-source speech and speaker-recognition toolkits on GitHub, implementing production-ready features and DevOps improvements that show a hands-on aptitude for backend and ML engineering. Based in Tianjin, he combines rigorous chemical engineering fundamentals with practical software skills to make technology transfers smoother and more measurable.
8 years of coding experience
Bachelor's degree, chemical engineering, Bachelor's degree, chemical engineering at Beijing University of Chemical Technology
English, Japanese, Chinese
Github Skills (25)
speaker-diarization10
pytorch10
python10
speaker-verification10
automatic-speech-recognition10
deep-learning10
speaker-recognition10
asr10
transformer10
speech-recognition10
c-language9
ml9
machine-learning9
trainings9
mle9
Programming languages (2)
Github contributions (5)
wenet-e2e/wespeaker
Mar 2022 - Dec 2022
Research and Production Oriented Speaker Verification, Recognition and Diarization Toolkit
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:75 reviews, 29 commits, 125 PRs in 9 months
Contributions summary:Chengdong primarily contributed to the project by adding features such as plotting DET curves, adding checkpoint functionality for model training, and supporting the setting of GPUs during embedding extraction. They also integrated a model averaging script. Furthermore, the user worked on enhancing the project's infrastructure by incorporating changes in the extraction process. Their work demonstrates a focus on improving model evaluation and simplifying model deployment.
pythonproduction-readyecapa-tdnnspeaker-verificationproduction
wenet-e2e/wenet
Apr 2022 - Oct 2022
Production First and Production Ready End-to-End Speech Recognition Toolkit
Role in this project:Back-end Developer / ML Engineer
Back-end Developer / ML EngineerContributions:5 reviews, 3 commits, 13 PRs in 6 months
Contributions summary:Chengdong's contributions focused on enhancing the WeNet toolkit, primarily through the addition of new recipes, bug fixes and refactoring existing code. They implemented context biasing with an AC automaton, contributing to more accurate speech recognition. They supported resampling features within the model, and upgraded the libtorch version. The user also addressed issues related to uncertain output results and fixed linting errors across different modules.
pytorchend-to-endasrproduction-readyspeech-to-text
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.