MELODY YIN — PERSONAL WEBSITE Agent-readable plain-text version of every written page on this site, generated for AI agents/crawlers. ================================================================================ HOME ================================================================================ melody yin (The homepage is an animated illustration of a pomegranate being cut open, with links to each section below appearing as the fruit bursts apart.) Section links: about, experience, projects, hobbies, contact ================================================================================ ABOUT ================================================================================ Hi, my name is Melody! I'm a junior at the USC Viterbi School of Engineering studying computer science and applied math, originally from SF Bay Area, CA. I have a huge range of interests in the field of computer science, ranging from AI for cybersecurity, to traditional SWE, to biomedical research! I'm currently learning a little bit of everything but am focusing primarily on the intersection of machine learning and systems: developing benchmarks for evaluating LLMs as security threats, building out my own startup in the art consulting space, and working hard on projects at IBM! Starter pack facts: - based in: bay area & la - studying: computer science - favorite movies: across the spiderverse & lord of the rings - currently watching: suits - probably drinking: an iced honey double shot latte ================================================================================ EXPERIENCE ================================================================================ --- Work / Research Experience --- IBM — Software Engineer Intern Durham, NC · May 2026 - Present Designed and built a scalable LLM-based system to automatically flag high-priority repository issues and PRs, streamlining the GitHub code review process; applied in Keycloak (6,000+ contributors, >2.5k issues). Artery — Solo Founder Los Angeles, CA · Jan 2026 - Present Designed and built Artery, an AI-driven hospitality art sourcing platform connecting interior designers at lifestyle hotels with vetted artists through intelligent brief-driven matching. Built the full-stack architecture, interactive prototypes, and pitch from scratch. Backed by Microsoft for Startups. INK Research Lab, USC — Research Assistant, URAP Awardee Los Angeles, CA · Aug 2025 - Present Proposed and led an independent project benchmarking the exploitative capabilities of out-of-the-box open-source models; explored synthetic distillation and post-training strategies for vulnerability exploitation research. ORBES — Embedded Systems Engineer Intern Los Angeles, CA · Jan 2026 - May 2026 Developed embedded software (implemented SLAM, integrated camera gyroscope/speedometer data for pathing) and a front-end dashboard for autonomous space robots used in industrial-scale space operations. Laboratory for ML, Health & Biomedicine, USC — Research Assistant, URAP Awardee Los Angeles, CA · Aug 2025 - May 2026 Developed deep learning-based methods to predict pMHC-TCR binding using molecular structure data. Su Lab, Scripps Research Institute — Data Science Intern La Jolla, CA · Jun 2024 - Dec 2024 Characterized aging-associated decay of transcriptional order and established relationships between gene set regulatory breakdowns and age. Presented at the 10th Aging & Gerontology Conference in San Francisco. Kundaje Lab, Stanford University — Machine Learning Intern Stanford, CA · Feb 2022 - May 2024 Designed and built a novel bimodal deep-learning model predicting gene expression from integrated data types, improving the lab's state-of-the-art model's performance by up to 15% in certain cell types. 2nd-author paper in peer review at Nature Genetics (preprint May 2024); presented at the 2023 Sigma Xi IFoRE conference. --- Education --- USC Viterbi School of Engineering B.S. in Computer Science · Expected May 2028 Relevant coursework: Object Oriented Programming, Algorithms, Discrete Math, Data Structures, Embedded Systems, Principles of Software Development, Linear Algebra, Probability Theory, Calculus III Awards: Viterbi Scholar, National Merit Scholar, URAP (Undergraduate Research Associates Program), Viterbi Dean's List The Harker School High School Diploma · Class of 2026 Relevant coursework: AP CS with Data Structures, Computer Architecture, Neural Networks, Expert Systems, AP Calculus BC, Multivariate Calculus, Differential Equations, Linear Algebra, AP Statistics --- Activities --- - Y Combinator Startup School (Jul 2026) - TroyLabs (USC Premier Startup Accelerator), Tech Team (Aug 2025 - Present) - Sigma Eta Pi, Omega Class Member (Jan 2026 - Present) --- Skills --- Languages & Programming Foundations: Go, Python, JavaScript, TypeScript, SQL, PostgreSQL, HTML, CSS, JSX, YAML, Bash, Shell Scripting Full-Stack Web Development: React, Next.js, Node.js, Express, Go Fiber, REST API, TailwindCSS, Radix, shadcn, Prisma, Supabase, Zod, PptxGenJS, Vitest, Playwright, Docker, Nginx, GitHub API AI, Machine Learning & Data Systems: Claude SDK, Azure OpenAI API, Ollama, RAG, Vector Embeddings, ChromaDB, Azure Cognitive Search, Prompt Engineering, CNN Design, Model Interpretability, Statistical Modeling, PCA, Data Pipeline Architecture Robotics, Computer Vision & Systems Infrastructure: SLAM, RTAB-Map, ROS 2, Sensor Fusion, IMU Integration, pyrealsense2, OpenCV, Point Cloud Processing, Redis, Asynq Miscellaneous: DeepSHAP, TF-MODISCO, Google Custom Search API, Web Scraping, Playwright Automation, Apify, PRO-cap, RAMPAGE, CAGE, ChIP-seq, ATAC-seq, DNase-seq, NGS Read Alignment ================================================================================ PROJECTS ================================================================================ Project PIE — IBM A scalable LLM pipeline that automatically triages high-priority issues and PRs to cut through code-review backlog, deployed on Keycloak — a repo with 6,000+ contributors and 2.5k+ open issues. Skills: Go, GitHub API, Prompt Engineering, Vector Embeddings, RAG, Data Pipeline Architecture Artery — Startup · Solo Founder An AI-driven art sourcing platform matching interior designers at lifestyle hotels with vetted artists from a creative brief. Designed and shipped the full-stack product, prototypes, and pitch solo — backed by Microsoft for Startups. Skills: React, Next.js, Node.js, TypeScript, PostgreSQL, TailwindCSS ORB — Orbes An embedded SLAM system fusing gyroscope and speedometer data for autonomous space-robot navigation, paired with a front-end dashboard for industrial-scale operations. Skills: SLAM, Sensor Fusion, IMU Integration, ROS 2, RTAB-Map, Point Cloud Processing ProCapNet and BPNet — Kundaje Lab, Stanford University Bimodal deep-learning models (built on BPNet) that predict transcription initiation directly from DNA sequence, uncovering the cis-regulatory motifs and syntax governing where and how often transcription begins — lifting the lab's prior state-of-the-art by up to 15% in select cell types. Paper: https://pmc.ncbi.nlm.nih.gov/articles/PMC11160661/ Skills: Python, CNN Design, Model Interpretability, DeepSHAP, TF-MODISCO, PRO-cap pMHC-TCR Binding Model — Laboratory for ML, Health & Biomedicine, USC Deep-learning methods that predict pMHC-TCR binding directly from molecular structure data, contributing to the lab's immune-response modeling research. Skills: Python, CNN Design, Statistical Modeling, PCA, Model Interpretability, Data Pipeline Architecture Swatch — Hackathon Project An automation platform for interior designers that pulls product data from vendor sites, organizes it into structured datasets, and generates client-ready slide decks in seconds, turning hours of data entry and formatting into a one-click workflow. Skills: Web Scraping, Playwright Automation, PptxGenJS, Node.js, TypeScript, Data Pipeline Architecture ================================================================================ HOBBIES ================================================================================ Things I do when I'm not at a keyboard — mostly drawing. All the art on this site (pomegranate included) is mine. From my sketchbook: a personal gallery of sketchbook drawings and a pencil portrait sketch. Museum favorites: pieces I keep thinking about after leaving the gallery. Everything else: reading Historical and anthropological books that teach me something I didn't know before — Nexus by Yuval Noah Harari is a recent favorite. cooking Experimental home cooking with a soft spot for Chinese and Korean flavors. Pomegranates show up in the kitchen as often as they do everywhere else. film I'll watch almost anything, but I have a soft spot for romcoms from the 2000s and The Lord of the Rings movies. gym & swimming Lifting and swimming laps — the reset button between classes and code. art & museums Contemporary art and museum-hopping whenever I travel — the Legion of Honor and the Brooklyn Museum are my favorites. ================================================================================ CONTACT ================================================================================ Based in San Francisco, CA / Los Angeles, CA. Always happy to chat! - Email: myyin@usc.edu - LinkedIn: https://www.linkedin.com/in/melody-y-yin/ - GitHub: https://github.com/myin25 ================================================================================ END OF FILE ================================================================================