Yan L. Email & Phone Number

Singapore

Singapore

> Popular Product PredictorTo maximize revenue in Livestream Sales Event, developed a model to assist the initial selection of top- performing products. It overcomes manual decision-making challenges and ensures high efficacy/accuracy. Scraped thousands of product information from Amazon; Executed EDA to clean data and extract features. Trained and optimized various classification models, including Logistic Regression, Decision Trees, Random Forests, Gradient Boosting, XGBoost, SVM… Show more > Popular Product PredictorTo maximize revenue in Livestream Sales Event, developed a model to assist the initial selection of top- performing products. It overcomes manual decision-making challenges and ensures high efficacy/accuracy. Scraped thousands of product information from Amazon; Executed EDA to clean data and extract features. Trained and optimized various classification models, including Logistic Regression, Decision Trees, Random Forests, Gradient Boosting, XGBoost, SVM, k-NN, and Ensemble Methods to combine some of them.Created a Flask-based web application for users to upload data, process it through the model, and view/download results. Deployed the app on AWS for accessibility. In all the models, XGBoost achieved the highest performance, with a recall score of 70% while maintaining precision. This improvement is estimated to save human resource costs ~120K annually.Github: https://github.com/Yyisangry/Popular_Product_Predictor_Repo Show less

Singapore

> Data Scientist_Patients SegmentationTo reduce patients undergoing traditional diagnosis which is invasive and risky, and accelerate patient selection in clinical trials, built a Patient Segmentation based on classification models. Collected clinical data with thousands of features, executed EDA, conducted Feature Engineering to reduce its dimensionality and selected important features. Successfully trained and optimized multiple classification models (Logistic Regression, Random… Show more > Data Scientist_Patients SegmentationTo reduce patients undergoing traditional diagnosis which is invasive and risky, and accelerate patient selection in clinical trials, built a Patient Segmentation based on classification models. Collected clinical data with thousands of features, executed EDA, conducted Feature Engineering to reduce its dimensionality and selected important features. Successfully trained and optimized multiple classification models (Logistic Regression, Random Forest, XGboost, SVM).Found the XGboost model performing best. Compared with the present best pre-screening, the precision score was improved by 35%. Benefiting from it, ~30% of patients would not suffer from anxiety and pain. It accelerated patient recruitment during clinical trials, and saved S$2.2 million annually in estimation.> Research Scientists_Pharmacology DepartmentTarget selection and MOA investigation: shortlist target candidates by comparative profiling andanalysis of multiomic data; develop in vitro or in vivo platform to evaluate pharmacodynamic response and efficacy; propose the potential MOA through analysis of open database or in-house experimental transcriptomic data, construct validation platform.Project management: document the detail project plan and timeline, cooperate and communicate with different functional department as team leader, present regularly within team or facing upper manager to report progress and make decisions at key time point. Show less

Wuxi, Jiangsu, China

> NLP Aided Systemic Literature ReviewerTo understand the research landscape of an unfamiliar area, anti-aging, in a short term with breadth and depth, developed a NLP aided Systemic Literature Reviewer.Performed systematic retraceable searches on databases (PubMed, Google Scholar, etc), collected abstracts of related literature, extracted key information and grouped them by some categories (such as, genes related to aging status, frequency appeared in publication, importance based… Show more > NLP Aided Systemic Literature ReviewerTo understand the research landscape of an unfamiliar area, anti-aging, in a short term with breadth and depth, developed a NLP aided Systemic Literature Reviewer.Performed systematic retraceable searches on databases (PubMed, Google Scholar, etc), collected abstracts of related literature, extracted key information and grouped them by some categories (such as, genes related to aging status, frequency appeared in publication, importance based on journal quality, etc) using NLP tool, TfidfVectorizer.Generated a comprehensive summary report within 3 weeks (3-fold shorter than manual literature review), and mainly contributed to the generation of a research proposal which received the CNY 300.000 grant support from Chinese Government (National Natural Science Foundation of China, No. 52103165).> Teaching and mentoringLecturer of Basic Biotechnology for undergraduates with students' evaluation ranking at top 15%. A second prize was awarded in the 2021 Teaching Competition.Guide the research for 3 postgraduate students, with one publication and industry delivery.Class advisor for a group of 32 undergraduate students, with positive feedback more than 90% students. Show less

Singapore, Singapore

> Protein engineering to optimize the enzyme industrial performance:Established protocols to construct mutation libraries using error-prone PCR and gene shuffling technologies, to enrich the mutation library size.Successfully constructed the plate-based high-throughput screening methods to avoid missing good mutants.> Insight Engine for Enzyme OptimizationTo reduce the time and labor spent on the traditional way to optimize enzyme, which is an important catalyst… Show more > Protein engineering to optimize the enzyme industrial performance:Established protocols to construct mutation libraries using error-prone PCR and gene shuffling technologies, to enrich the mutation library size.Successfully constructed the plate-based high-throughput screening methods to avoid missing good mutants.> Insight Engine for Enzyme OptimizationTo reduce the time and labor spent on the traditional way to optimize enzyme, which is an important catalyst applied in the industry production, conducted Insight Engine based on the decision tree model.Generated a random library of enzymes by experiments; encoded enzyme properties as features, and labeled the enzyme performance by experimental result; then used these data to train the decision tree based model. Finally, confirmed some key features which are important for the desirable enzyme performance.It significantly reduced the feature numbers for further optimization which can only be verified by lab experiments. As such, the experimental selection pool was reduced by 16.7%. Only in terms of labor cost, it saved S$84, 000 annually. Show less

Doctor Of Philosophy - Phd, Bioengineering And Biomedical Engineering

Activities and Societies: Global Young Scientist Summit (GYSS, with multiple Nobel Prize Winners), poster presentation, 2019. >>.

Master'S Degree, Biochemical Engineering

Protein Engineering to improve enzyme activity through directed evolution (cooperated with Yihai Kerry). >> Comprehensive experience.

Bachelor'S Degree, Advertising

Exploration of the effect of nitrogen source type on bio-synthesis of DHA >> Culture and analyze the DHA production in S. limacinum under.