Available for 2026 Roles

Zhiyuan Song

Senior, B.S. in Computer Science

University of California, Santa Cruz
Passionate about Full-Stack Development, AI Applications, and Scalable Systems

2
Nature Co-authored
6+
Open Source Projects
3.6
GPA (4.0 Scale)
Full-Stack
Next.js & Python

Publications

Published

MRI-based Multi-organ Clocks for Healthy Aging and Disease Assessment

H. Cao*, Z. Song*, M.R. Duggan, G. Erus, D. Srinivasan, Y.E. Tian, W. Bai, et al.
Nature Medicine, 2025  |  doi: 10.1038/s41591-025-03999-8

Read Abstract

We developed 7 organ-specific MRI-based biological age gaps (MRIBAGs) for brain, heart, liver, adipose tissue, spleen, kidney, and pancreas. Leveraging imaging, genetic, proteomic, and metabolomic data from 313,645 individuals, we linked these MRIBAGs to 2,923 plasma proteins, 327 metabolites, and 6.5M genetic variants. GWAS identified 53 significant locus pairs, and we prioritized 9 druggable genes as potential anti-aging targets. The MRIBAGs demonstrated clinical relevance in predicting disease endpoints, mortality, and cognitive decline trajectories.

Under Review

Sleep Chart of Biological Aging Clocks Across Organs and Omics

C.K. O'Toole, Z. Song, F. Anagnostakis, Z. Yang, Y.E. Tian, M.R. Duggan, et al.
Submitted to Nature, 2025  |  medRxiv preprint

Read Abstract

We propose a Sleep Chart assessing the relationship between sleep duration and 23 biological aging clocks across 17 organ systems and 3 omics types. A systemic U-shaped pattern shows both short (<6h) and long (>8h) sleep linked to elevated biological age gaps, with optimal sleep varying by organ and sex (6.4-7.8 hours). Abnormal sleep duration is consistently linked to increased risk of systemic diseases and mortality. Short and long sleep affect late-life depression via distinct pathways. Our findings suggest the U-shaped relationship is driven by modifiable sleep disturbances, highlighting sleep optimization potential for healthy aging.

Professional Experience

Full-Stack Engineer @ Loop Interview

Apr 2025 - July 2025
  • Developed and launched a paid AI interview platform with customizable sessions and real-time feedback.
  • Built the AI engine using Gemini API and prompt engineering to enable multi-turn, multi-role, tone-adaptive interviews.
  • Created personalized interview generation by extracting key info from resumes and JD using regex and LLMs.
  • Integrated ElevenLabs for voice synthesis and Three.js for visual interviewer presentation.
  • TECH: Next.js · TypeScript · Flask · Firebase · Docker · GCP

Web Developer Intern @ Columbia University (Lab of AI & BioMed)

Dec 2024 - July 2025
  • Co-authored two papers (Nature & Nature Medicine) for contributions to research platform development.
  • Refactored lab website architecture and built analytics charts with D3.js, handling 200+ pages.
  • Boosted platform traffic by 5,000+ visits post-launch.

Tech Lead & Full-Stack Developer @ 3CCA Church

Jan 2025 - June 2025
  • Led the development of a new online platform using the MERN stack (MongoDB, Express.js, React, Node.js).
  • Built the admin system and core modules (events, audio, articles) to support non-technical staff operations.
  • Designed system architecture and flexible MongoDB schemas to enable scalable content management.

Featured Projects

TalentSync

An AI-powered job platform to solve resume-JD mismatch and weak cover letter alignment. Features automated resume extraction, JD analysis with HR/HM/ATS scoring, and personalized cover letter generation.

MERN Gemini API Stripe Redux

Study Assistant

AI-driven learning platform that transforms educational materials into personalized study content and mock exams. Features smart document processing, AI content generation, and semantic search with pgvector.

Next.js TypeScript PostgreSQL pgvector OpenAI

AnimeHub

A multilingual anime community platform for sharing and tracking collections. Features user rating, filtering, secure authentication, and modular backend architecture.

MongoDB Express React Node.js Redis

Personal Website Template

A flexible and highly customizable personal website template system for job seekers and researchers. Features interactive onboarding, skills showcase, education timeline, and easy Vercel deployment. 100+ downloads with 95% deployment success rate.

Next.js React TypeScript Tailwind Vercel

WordRange AVL

Optimized data structure using self-balancing AVL tree for efficient insertion and range queries on large-scale text data. Handles up to 2 million operations per minute with logarithmic time complexity.

C++ AVL Tree Algorithms

WebCrawl

A modular web crawler implemented in C. Features URL fetching, HTML parsing, hashtable for visited URLs, and page storage. Uses libcurl for HTTP requests with configurable crawl depth.

C libcurl Data Structures

Technical Proficiency

Languages

  • JavaScript / TypeScript
  • Python
  • Java
  • C / C++
  • SQL

Frontend

  • React.js / Next.js
  • Tailwind CSS
  • Redux
  • D3.js

Backend & Ops

  • Node.js / Express
  • Flask
  • MongoDB / PostgreSQL
  • Docker / Google Cloud

AI & Tools

  • Gemini / OpenAI API
  • Prompt Engineering
  • ElevenLabs
  • Git / GitHub Actions