Shray V. Email & Phone Number

• Ongoing Research: Generation of biologically significant protein structure confirmations• Explored slow feature analysis dimensionality reduction on molecular simulation data for identification of biologically significant features using python, numpy, and pandas• Generated collective variable for plasmepsin-II for metadynamics, successfully accelerating a cryptic pocketopening• First author on paper accepted by PLOSOne

Sacramento, California, Us

• Fine-tuned various natural language models for inference of business identification numbers using noisy business data• Scripted AWS Glue ETL Jobs for flattening and cataloging of business data into data lakes

• Developed Transformer-based machine learning architecture for Enzyme Thermostability prediction using structure-aware protein embeddings for undergraduate thesis• Developed dataset pipeline for labeling of Protein Interface residues using gemmi and python to train machine learning models for Protein Protein Interface identification to prevent Protein Aggregation

New York, Us

• Implemented a tool for adaptive sampling of molecular dynamics simulations using k-means clusteringand Markov State Models to find potential new conformation states for therapeutic binding targets usingpython, mdtraj, and deeptime• Tested adaptive sampling pipeline on a potential therapeutic target protein

Boston, Ma, Us

-Developed C++ data readers and processors for 3 and 4 dimensional CSV file representation of mass spectrometry scan data-Designed a layered data structure and accompanying algorithm in order to slice data across 3 dimensions as optimally as possible using the Qt library-Started linking CSV file readers with plot and heatmap GUI elements in order to visualize different segments of data

Boston, Ma, Us

-Scraped terabytes of company data using the Dropbox API, Python, and SQLite to train machine learning models, and index all company files-Used Python to write an annotator to extract the fragmentation pattern of peptides which have gone through mass spectrometers using Scikitlearn and Pandas-Trained machine learning models using Tensorflow and Keras in order to improve upon existing peptide fragmentation prediction models

Baltimore, Md, Us

- https://netzarroyo.jh.edu/- The Netz Lab is a pharmacology lab at Johns Hopkins University's School of Medicine that focuses on the improvement of electrochemistry-based biomedical sensors in order for them to detect significant molecules in-vivo in real time.- Improved upon existing real-time data analysis code by optimizing processing and calculations using multiple types of data, and further adding to the data visualization aspect of the pre-existing software.-Created Python-based Graphical UI for an open-source potentiostat to be fused with current software in order to collect and analyze in-vivo data harmoniously and more efficiently

San Jose, California, Us

-https://contiq.com-Software Development Company focusing on increasing efficiency and win rate of sales managers and marketers using Artificial Intelligence and Machine Learning-Contributed new functionalities to the product in the form of image comparisons, initially pixel by pixel and now by defining and highlighting key differences between PowerPoint slides-Used Optical Character Recognition to determine whether any changes had been made to the text on slides-Performed quality testing for the product by participating in various bug bashes by recording and assigning any errors found

Doctor Of Philosophy - Phd, Bioengineering And Biomedical Engineering

Bachelor'S Degree, Computer Science

Education record