Jason Hu

Menlo Park, CA
jasonyhu@gmail.com

Education

Purdue University
B.S. Data Science & Computer Science, Double Major
Expected 2027
Relevant coursework: CS 348 (Information Systems), CS 373 (Data Mining & Machine Learning), CS 251 (Data Structures & Algorithms), STAT 416 (Probability), CS 240 (Programming in C), CS 250 (Computer Architecture), CS 252 (Systems Programming)

Skills

Cloud & DevOps
AWS (S3, CloudFront, Lambda, API Gateway, DynamoDB, IAM), AWS SAM, CloudFormation, Infrastructure as Code, CI/CD, GitHub Actions, Git, pytest
Programming
Python, Java, C, R
Tools
MATLAB, TensorFlow, Keras, NumPy, Azure Databricks, Torchvision
Salesforce
Declarative Development, Data Security, Reports/Dashboards
Strong team player with the ability to work independently, learn quickly, and pay attention to detail. Excellent oral and written communication skills.

Certifications

CompTIA Network+

Projects

Cloud Resume Challenge — Serverless AWS Portfolio Site
2026
AWS · Python · Infrastructure as Code · CI/CD
github.com/zJasonH/cloud-resume-backend
  • Deployed a static resume site to S3 with CloudFront as a CDN, served over HTTPS with a bucket policy scoped to public object reads
  • Built a serverless visitor-counter API using API Gateway (HTTP API), Lambda (Python 3.12), and DynamoDB, applying atomic ADD updates to prevent race conditions under concurrent requests
  • Migrated console-provisioned resources to infrastructure as code with an AWS SAM template, replacing account-wide AmazonDynamoDBFullAccess with a table-scoped least-privilege IAM policy
  • Configured CORS at the API Gateway layer and managed deployments through CloudFormation changesets with retain policies on stateful resources
  • Built a CI/CD pipeline with GitHub Actions that runs pytest unit tests against a mocked DynamoDB backend on every push, gating the SAM build and deploy so only passing code reaches AWS
  • Automated frontend releases with a second workflow that syncs the site to S3 and issues a CloudFront invalidation, authenticating via GitHub OIDC federation so no long-lived AWS credentials are stored in source control

Experience

Cyber Space Security Lab — Dr. Elisa Bertino, Purdue University
Sep 2024 – Present
Lab Intern (Adversarial ML)
  • Benchmarked adversarial robustness of ViT-B/16 across 90 adapted models — 3 datasets × 6 adaptation methods (linear probing, adapters, prefix/prompt tuning, LoRA, full fine-tuning) × 5 training objectives — under 6 black-box transfer attacks from PiT-B/DeiT-B surrogates, 3 ViT-specific white-box attacks, and AutoAttack
  • Quantified the robustness gap in parameter-efficient fine-tuning: standard adaptation holds above 93% clean accuracy while black-box robust accuracy stays below 18%, and PGD drives white-box robust accuracy to near zero at every learnable-parameter budget
  • Co-author on STAR (Smoothness and Token Alignment Regularization), under review at ICLR 2027 — a regularization method reaching ~72% black-box robust accuracy on CIFAR-10 under LoRA with 27 points higher clean accuracy than PGD adversarial training, and +30 points robust accuracy over MixAT on CIFAR-100, at cost comparable to adversarial training
  • Contributed method, architecture, and related-work sections and produced pipeline and threat-model diagrams for the submission
Lab of AI in Medicine and Biomedical Physics — Dr. Lei Xing, Stanford University
May – Aug 2024
Academic Intern (AI in Medicine)
  • Benchmarked and optimized gene feature extraction methods for efficiency
  • Optimized methods to improve computational large-scale medical AI applications
The Data Mine, Purdue University — with BASF
Aug 2023 – Apr 2024
Research Intern
  • Built Naive Bayes models to identify market opportunities from large-scale datasets
  • Utilized Databricks for data processing and Tableau to create visual dashboards, presenting findings and actionable insights to project mentors and BASF stakeholders
Stanford University — Statistics Department
Jun 2022 – May 2023
Academic Intern (Computational Biology)
  • Performed scRNA-seq data analysis with PCA, t-SNE, UMAP, and clustering in Python/R
FedEx
Summer 2026
Lead Document Sorter
  • Led sort operations for documents, small packages, and parcels in a time-critical distribution environment, directing workflow to meet fixed departure deadlines under sustained high volume
  • Triaged and prioritized time-sensitive shipments, making rapid routing decisions to ensure priority freight was staged and dispatched to the correct destinations on schedule
  • Maintained sorting accuracy under tight cutoff windows, identifying and correcting misroutes before dispatch to prevent downstream delivery failures
Apprentice Education
Dec 2022 – 2023
Volunteer Salesforce Administrator/Developer
  • Built custom CRM features, reports, and dashboards to manage student engagement
  • Implemented Salesforce tools to improve workflows and support data management needs
Loading visitor count…