Sai Mahit V. Email & Phone Number

West Lafayette, In, Us

Cambridge, Massachusetts, United States

- Collaborated with a team of computational and medicinal chemists on developing computational tools that will help find potent rare-disease drugs- Retooled the REINVENT reinforcement learning (RL) algorithm for use in company-related drug discovery pipelines by implementing 2D structural similarities along with 3D Shape, Pharmacophoric, and Electrostatic Potential based similarity scoring. Generated molecules were further analyzed using cheminformatic descriptors, ADME properties, Protein Docking, and Retrosynthetic Feasibility- Evaluated the performance of AI-driven retrosynthetic tools with the help of other computational and medicinal chemists- Embedded Machine Learning (ML) based conformational sampling tools to generate possible conformers in the RL workflow- Advised the Data Science, IT, and Data Ops team in the day-to-day maintenance and development of an AWS GPU cluster

West Lafayette, Indiana, United States

- Developed edge-based graph attention networks (EGAT) to predict both reaction and molecular properties such as activation energies with a fellow graduate student. EGAT models can be 1) trained from scratch, 2) with pre-trained weights as a starting point, or 3) using ablation models- Implemented Unifold Mapping Approximation (UMAP) to visualize the learned reaction/molecular latent space from EGAT models - Benchmarked EGATs performance on 1) the RGD1 dataset and 2) various molecular datasets with neural networks, random forest, and graph convolutional networks- Examined both the transferability of trained models on external datasets and the effect of inherent error in quantum chemical calculations on trained models to facilitate informed decisions on database generation by testing the trained models on external datasets and performing further validation of the computational data generated - Processed in-house data on free radical and ion creation and benchmarked them using external datasets such as NIST and ALFABET- Incorporated methods to explore bond cleavage into the Yet Another Reaction Exploration (YARP) package to 1) atom-map bond dissociation and protonation and 2) search for possible Electron-Ionized Mass Spectra (EI-MS) fragments from a given starting molecule such that EGAT can predict EI-MS peaks given a molecule and fragment- Initialized the development of EGAT models that 1) can predict atom and/or bond-level properties for a given molecule, 2) implement added machine learning tools for improved prediction- Developed tools that utilize PyVis to interactively visualize reaction networks- Contributed to the creation of the Reaction Graph Depth 1 (RGD1) Dataset comprising of 200,000 high-level and 700,000 low-level transition state geometries of model reactions- Prepared onboarding lectures on the RGD1 dataset and EGAT so that others in the group can utilize the methods for their research

West Lafayette, Indiana, United States

- Provided support to 2nd and 3rd-year chemical engineering students in the undergraduate fluid mechanics course taught by Professors Kendall Thomson and Michael Harris- Instructed students in utilizing Python for solving day-to-day problems in fluids- Taught recitation sections, designed exam review sessions, and ran office hours to give students a deeper understanding of course material- Proctored and Graded exams to oversee student performance- Collaborated with another TA and Lab Manager to oversee the teaching of the Fundamentals Lab

West Lafayette, Indiana, United States

- Worked on developing a set of synthetic data generation tools through the Purdue Data Mine- Led two groups of eight people from doctorate, masters, and undergrduate backgrounds based both in Purdue and Arizona State (ASU) that analyzed trends, important genes, and uncertainty in Alzheimer’s disease data using statistical analysis.- Conducted RNA transcription data simulation using Monte Carlo and Latin Hypercube sampling based on a multivariate distribution, studying outlier impact on Alzheimer's Disease patient classification.- Created a cell-free classifier analyzing mRNA to detect potential Alzheimer's disease presence in patients

West Lafayette, Indiana, United States

- Provided support to 2nd and 3rd-year chemical engineering students in comprehending statistical models for CHE 320 course, led by Professors Kendall Thomson and Can Li- Instructed students in utilizing JMP software for analyzing data correlations using ANOVA, emphasizing statistical techniques on diverse datasets- Designed exam review sessions and office hours to help students improve their understanding of the material

Newark, Delaware, United States

- Developed a Genetic Algorithm in Python that determines nanoparticle models in two ways: 1) based on both Infrared (IR) and Extended X-ray Absorption Fine Structure (EXAFS) spectra and 2) based on the radial distribution function and coordination environment - Investigated the effects of substitutional vacancies and molecular dynamics on nanoparticles' geometric and spectroscopic properties- Presented the use of theoretical spectroscopy modeling of EXAFS to others at group meetings and research symposia- Authored an undergraduate thesis on the integration of Genetic Algorithms to develop spectra-based computer nanoparticle 3D models using multiple spectroscopies

Newark, Delaware, United States

- Assisted 1st-year science/engineering students in the Introduction to Computer Science and Programming (CISC 106) course with Dr. Katie Wassil- Assisted students towards gaining a better understanding of Python and MATLAB coding through group and 1-on-1 help during lab sessions- Proctored online exams with fellow Teaching Assistants and Instructors

Newark, Delaware, United States

- Evaluated Homework Set problems for 1st-year chemical engineering students taking the Introduction to Chemical Engineering course (CHEG 112)- Reported student scores to graduate teaching assistants

Mechanicsburg,Pa

- Tutored willing students from middle and high school on Math, Science, and English - Helped student prepare for the SAT

Mechanicsburg,Pa 17050

- Taught basic geography skills and helped coordinate Geography Bees with faculty and PTO at 2 schools - Promoted the teaching of geography and geography bees across school district

Bethlehem, Pennsylvania, United States

- Collaborated with a group of fellow high schoolers to write a Technical Review of the Hyperloop technologies and their potential viability in the next 5-10 years- Designed a Microfluidic reactor device that utilized a sulfuric acid catalyst to ensure full reactivity of Potassium Iodate with water - Developed possible plans to viably remediate the marshy habitat around abandoned coal mine effluent tunnel in Lehigh Gorge State Park in Jim Thorpe, PA in conjunction with the Sierra Club- Designed a 3D printed I beam and benchmarked its strength and deformation using Solidworks

Doctor Of Philosophy - Phd, Chemical Engineering

Activities and Societies: Tau Beta Pi (Graduate Advisor), Murdock Science Club

Bachelors In Chemical Engineering (B.Ch.E.), Chemical Engineering, 3.76

Activities and Societies: Telkes Scholar, Honors Program, Tau Beta PiMinors: International Business with Spanish, Materials Science.

Diploma, 94/100 Unweighted

Activities and Societies: Quiz Bowl, Technology Student Association, Moody's Mega Math Competition, Odyssey of the Mind, LifeSmarts, Key.