Skip to content
View Code-Tomato's full-sized avatar
5 followers · 1 following

Achievements

Achievement: YOLOAchievement: Pair ExtraordinaireAchievement: Pull Shark

Achievements

Achievement: YOLOAchievement: Pair ExtraordinaireAchievement: Pull Shark

Highlights

  • Pro

Block or report Code-Tomato

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.

Content in all repositories owned by your account will be closed.
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
Code-Tomato/README.md
tomato mark
Typing SVG

LinkedIn Portfolio Email

👋 Hi, I'm Nathan Lemma · ECE @ UT Austin · Systems Engineer

🔬 Now: Agentic ML systems • Software Engineering Intern @ Caterpillar

⏪ Previously: SysML Research Group • Longhorn Racing (Battery Management) • FAST Research Group

⭐️ Featured Projects

Project What it is
EnerTune Baselines Four published GPU-sharing serving systems rebuilt and ported to a 16-A100 cluster
MLLM-Energy 1,080 profiled configs showing a one-size GPU split wastes energy across multimodal LLM phases
CHT-Radix Seven concurrent hash tables/radix trees benchmarked against a real ShareGPT trace
Pintos Teaching kernel extended with user processes, syscalls, demand paging, full filesystem

📄 Publications

"Beyond Utilization: Energy-Conscious GPU Sharing for Inference Serving" SOSP '26 · 3rd author Rebuilt and ported the four GPU-sharing serving systems EnerTune is benchmarked against onto a 16-A100 cluster — the evaluation behind its 1.4–2.3x energy reduction.

Pinned Loading

  1. Code-Tomato.github.io Code-Tomato.github.io Public

    Portfolio Website

    Astro

  2. CHT_Radix CHT_Radix Public

    C++17 concurrent hash tables and radix trees, benchmarked on synthetic and LLM-derived prefix-caching workloads.

    C++ 1 1

  3. sus-gpus-baselines sus-gpus-baselines Public

    The collection of all the baselines we used

  4. mllm-energy/MLLM-Energy mllm-energy/MLLM-Energy Public

    Composite result of Spring 2026

    HTML