Tilly Seesillapachai Email & Phone Number

Boston, Ma, Us

Cambridge, Ma, Us

Cambridge, Massachusetts, United States

Boston, Massachusetts, United States

• Developed supervised machine learning models (SVR, XGBoost, RF, ANN) to predict compound logKp (skin permeation coefficient) for ranking and screening during lead discovery and optimization• Secured additional project funding by demonstrating the effectiveness of the prototype model in ranking compound permeability, ensuring the project's continuation• Created a custom data loader in PyTorch to facilitate training of AbbVie’s internal GNN model, enabling researchers to add new endpoints• Tested Graph Neural Network (GNN) capabilities to address the small data regime of logKp datasets• Added global features to node features in AbbVie’s internal GNN model

Chicago, Illinois, United States

• Designed a multi-task, domain-adversarial machine learning model in TensorFlow for survival analysis of breast cancer datasets• Utilized Ray Tune for hyperparameter tuning of the multi-task, domain-adversarial survival analysis model• Performed exploratory data analysis (EDA) on clinical breast cancer datasets to understand their underlying distribution• Used TensorFlow to apply a Neural Additive Model for interpretable survival analysis• Displayed findings at the 2023 NCI Informatics Technology for Cancer Research (ITCR) Conference

Remote

• Applied the DeBERTa model and CLAMP, a clinical NLP toolkit, in PyTorch to filter self-reported COVID-19 and Monkeypox vaccination side effects from Twitter and Reddit • Analyzed over 550,000 tweets in GCP BigQuery using Python and SQL to compile the top 10 most frequently tweeted side effects• Created a U.S. heat map visualizing the frequency of self-reported side effects using Shiny in R• Displayed findings at the 2022 Observational Health Data Sciences and Informatics (OHDSI) Symposium

Evanston, Illinois, United States

• Held weekly office hours to assist 150+ undergraduate students with machine learning lecture and homework inquiries• Enhanced the data science curriculum by offering constructive feedback

Evanston, Illinois, United States

• Held weekly drop-in tutoring sessions answering questions regarding lectures and homework from single variable calculus, sequences and series, and linear algebra courses• Held weekly review sessions teaching general chemistry

Evanston, Illinois, United States

• Developed a novel bioinformatics tool for deducing peptide sequence motifs from atomic protein-peptide complex structures using BLAST, Clustal Omega, and WebLogo• Deployed this tool onto MONSTER, a web tool cited in 45+ research papers, used to catalog and visualize intramolecular interactions in proteins and macromolecular complexes• Updated the user interface of MONSTER using vue.js JavaScript framework