Skip to content
View tonyyunyang's full-sized avatar
💭
📚
💭
📚
6 followers · 25 following

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: Galaxy Brainx3Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: Galaxy Brainx3Achievement: YOLOAchievement: Quickdraw

Highlights

  • Pro

Block or report tonyyunyang

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
tonyyunyang/README.md

tony@amsterdam:~ $ whoami: Tony (Tongyun) Yang, Independent AI Researcher in Amsterdam. focus.txt: the scaling race is becoming a systems race; bigger models still matter but are not enough; what counts is how we train, route, compress, and ship them; building AI that is capable, efficient, and genuinely useful. open-to-work: left my PhD fellowship this spring, open to research roles in academia or industry.

site scholar email cv en cv zh

~/publications $ ls -lt: twinrouterbench (ACM CAIS 2026): agentic LLM routing benchmark, cuts cost 53%. mera (ACM CAIS 2026): trace-driven routing, 87% accuracy at about half cost. through-the-eyes-of-emotion (IMWUT 2025): eye-tracking dataset for emotion in VR. reverse-imaging (MICCAI 2025): any-sequence cardiac MRI segmentation. pruning-nnunet (MIDL 2025): over 80% of weights pruned, dice above 0.95.

twinrouterbench paper twinrouterbench code twinrouterbench site   mera paper mera code
through-the-eyes paper through-the-eyes code through-the-eyes data   reverse-imaging paper reverse-imaging code
pruning-nnunet paper pruning-nnunet code

~/projects $ ls -lt: llm-router: routing benchmark with Gradient Networks, under review at NeurIPS 2026. human-intent-world-model: a vision-language model that reads customer intent, with MeetaVista. polymarket-decoder: a self-driven order-book study of prediction markets.

llm-router code

~/stack $ cat stack.toml: daily: python, pytorch, cuda, latex. ai: claude code, codex, cursor, a custom harness (coming soon), pi-agent, harness. infra: vllm, sglang, ray, deepspeed. often: tensorflow, unity, typescript, c++, linux. shelf: react, node, docker, git.

~/log $ git log: a contribution heatmap with total, current streak, longest streak, and busiest day over the last year, regenerated twice a day by a GitHub Action.

~/life $ cat .life: from Sichuan, now Amsterdam; cooks in a Sichuan wok, never a covered pot; tennis with a pure drive and a 1:43:53 half-marathon; reads Sartre and Borges, tulips each spring; adopting a 狸花猫 cat named 瓜子.

~/contact $ ./contact.sh: email tonyyunyang@outlook.com, github.com/tonyyunyang, Google Scholar, tonyyunyang.github.io.

email github scholar site

plain text (for screen readers and search)

Tony (Tongyun) Yang, Independent AI Researcher in Amsterdam. The scaling race is becoming a systems race; bigger models still matter, but they are not enough; what counts is how we train, route, compress, and ship them. Left my PhD fellowship this spring; open to research roles in academia or industry.

Publications: TwinRouterBench (ACM CAIS 2026), MERA (ACM CAIS 2026), Through the Eyes of Emotion (IMWUT 2025), Reverse Imaging (MICCAI 2025), Pruning nnU-Net (MIDL 2025).

Projects: LLM Router (routing benchmark with Gradient Networks, under review at NeurIPS 2026), Human Intent World Model (vision-language model for customer intent, with MeetaVista), Polymarket Decoder (self-driven order-book study of prediction markets).

Stack: python, pytorch, cuda, latex; claude code, codex, cursor, a custom harness, pi-agent, harness; vllm, sglang, ray, deepspeed; tensorflow, unity, typescript, c++, linux; react, node, docker, git.

Contact: tonyyunyang@outlook.com, github.com/tonyyunyang, scholar.google.com, tonyyunyang.github.io.

hand-built, no template ancestry · companion to tonyyunyang.github.io

Pinned Loading

  1. CommonstackAI/CommonRouterBench CommonstackAI/CommonRouterBench Public

    Python 11

  2. Scholar-High-Lights Scholar-High-Lights Public

    JavaScript 2

  3. MultiRepEyeVR/Through-the-Eyes-of-Emotion MultiRepEyeVR/Through-the-Eyes-of-Emotion Public

    Python 5 3