Sriram Hathwar Applied math student at Princeton with strong application interests in computational biology and drug discovery.

About Me

I am a senior at Princeton studying applied math, with research interests in computational biology and genomics.
I enjoy working on technically challenging problems with high impact in biology and healthcare.

Experience

• Computational Biologist, RAPT Therapeutics, Aug 2022 - Jul 2024
  • Developed a tool in R to identify combination therapy partners and predictive biomarkers for leading cancer drug in clinical trials
  • Applied deconvolution algorithms to bulk RNAseq patient biopsy samples to identify cell type specific gene expression profiles
  • Saved discovery biology team $30K by demonstrating similarity in measurements between in-house assays and external studies
  • First authored internal poster, presented findings to Scientific Advisory Board


• Bioinformatics Research Assistant, Washington University School of Medicine , May - Aug 2021
  • Trained diffusion maps, graphical neural networks, and LSTM frameworks on single-cell RNA-seq data from a human embryo and gastruloid to model germ layer formation
  • Discovered potential heterogeneity in mesoderm formation
  • Intend to deploy model for public use and confirm results experimentally


• AI + Healthcare Instructor, Inspirit AI, Nov 2020 - Present
  • Teach high school students about applications of cutting edge AI techniques like graphical neural networks and transformers to healthcare
  • Explain math and theory behind various machine learning models including logistic regressions, neural networks, CNNs, RNNs, and LSTMs
  • Guide students throgh Jupyter notebook projects in Python including seizure prediction, skin cancer detection, mental health chatbots, and drug discovery


• Computational Neuroscience Assistant, Seung Lab, Nov 2020 - Apr 2021
  • Used k-means clustering as baseline model to cluster neuronal types in mice
  • Applied literature algorithms like time-inhomogeneous diffusion condensation to compare clustering outcomes with baseline model


• Co-founder and CEO, Angle Global, Jun 2019 - Jul 2021
  • Signed licensing contract to be exclusive distributor of English learning content to the 5 million participants of the Spelling Bee of China
  • Earned $25,000 grant to work on company over summer 2020 through Ovo Fund
  • Selected for Neo Hack Month and advanced to final interview stage for YC S20

Projects

• AI Flashcard Generator
  • Designed and developed a full-stack web application that leverages LLMs to generate flashcards from user-uploaded PDFs
  • Deployed web app using LLMs, vector databases, and RAG to generate flashcards with linked sources, built with Svelte frontend and Python FastAPI backend
• Integrated Latent Variable Model for Lapses in Rodent Decision-Making
  • Working under Princeton Professor Jonathan Pillow to model mice decision making in sensory analysis experiments for senior thesis
  • Exploring an integrated approach using reinforcement learning and Hidden Markov models to model stimulus-independent error rates in rodent perception tasks


• Replicating Large Scale Multi-Omic Analysis of COVID-19 Severity
  • Led undergraduate team to reproduce published results analyzing metabolites, proteins, transcripts, and lipids from COVID-19 patient blood samples
  • Wrote SQLite queries to access patient data from distinct tables and performed standard dimensionality reduction techniques like PCA for further analysis
  • Developed random forest machine learning model to predict patient COVID-19 status and severity based on multi-omic dataset from Overmyer et. al (2020)

Awards

Activities

Links