Melissa Pan (@melissapan) / X

Melissa Pan

955 posts

Melissa Pan

@melissapan

CS PhD @UCBerkeley Sky Lab 🐻 Systems & AI & Sustainability 🌍 Prev: @google, @ibm, @CarnegieMellon🐕‍🦺, @UofT🇨🇦

Berkeley

melissa-pan.github.io

Joined October 2023

Pinned
Melissa Pan
@melissapan
Jul 14
Checkout a preview for our latest work BRANE 🧠 on pushing the pareto frontier of retrieval agents! Less 💰 for the same or even higher quality!
Arena.ai
@arena
Jul 14
Right before joining Arena, @melissapan (PhD candidate, UC Berkeley) presented research on cutting agent system costs by 89% — while matching 100% of the best static config's accuracy. Picking the right LLM matters less than you think. Full system config >> LLM routing alone.
00:00
Melissa Pan
@melissapan
Jun 14, 2025
multi-agent outperforms single agent by 90.2% is very interesting. One reason we haven't seen multi-agents winning is that existing benchmarks are rather "simple." This makes multi-agents seem more like a PoC than a necessity, which is not a true reflection of MAS's capability.
Anthropic
@AnthropicAI
Jun 13, 2025
New on the Anthropic Engineering blog: how we built Claude’s research capabilities using multiple agents working in parallel. We share what worked, what didn't, and the engineering challenges along the way. anthropic.com/engineering/bu…
Melissa Pan
@melissapan
Sep 29, 2025
It’s PhD / grad school application season again🌱 I’ve been getting reach out about SoPs, research fit, “should I do PhD at all??” Here's the dump of my honest thoughts from my application experience - what worked for me, and how I decided if a PhD was right for me. Wrote it
Melissa Pan
@melissapan
Jan 2, 2025
2024 was an exciting year! excited to share that I - 1. Graduated from Carnegie Mellon 2. Spent the summer at Google working on edge TPU-GPU co-compilation 3. Started my PhD at UC Berkeley @BerkeleySky
Melissa Pan
@melissapan
Apr 25, 2025
🚨 Why Do Multi-Agent LLM Systems Fail? ⁉️ 🔥 Introducing MAST: The first multi-agent failure taxonomy - consists of 14 failure modes and 3 categories, generalizes for diverse multi-agent systems and tasks! Paper: arxiv.org/pdf/2503.13657 Code: github.com/multi-agent-sy… 🧵1/n
Melissa Pan
@melissapan
Oct 21, 2025
I'm very grateful and honoured to receive the Amazon AI Fellowship. Thank you @AmazonScience for the generous support in helping to drive our research forward! 💛🐻💙
UC Berkeley EECS
@Berkeley_EECS
Oct 21, 2025
Amazing! 10 @berkeleyeecs @SkyCompLab grad students are Amazon AI PhD Fellows! Congrats! Learn more about our fellows here: bit.ly/48Hl88q #AmazonAIFellowship @BerkeleySky
Melissa Pan
@melissapan
Sep 18, 2025
Excited to share: MAST has been accepted as 🌟 NeurIPS D&B Spotlight🌟 Updates for the community: - NEW: We open-source 1,000+ multi-agent traces (link in 🧵). - lots of exciting use cases are emerging, we’ll be releasing blogs & tutorials to help you get started - And … more
Melissa Pan
@melissapan
Oct 4, 2025
Today i was titled the ** only ** person in the lab active on twitter (by @pgasawa and @aczhu1326 ). Given my inactivity on twitter, im concern about our lab’s social media presence.
Melissa Pan
@melissapan
Oct 9, 2025
AI is going to replace researchers? 🙀 AI is going to replace PhD students? 🙀 AI is going to take all of our jobs? 🙀 NO! But… 🚨AI is upending systems research 🚨 We show that by leveraging AI-driven research systems (ADRS), we can drastically accelerate the algorithm
00:00
AI-Driven Research for Systems
@ai4research_ucb
Oct 9, 2025
🚀 Excited to release our new paper: “Barbarians at the Gate: How AI is Upending Systems Research” We show how AI-Driven Research for Systems (ADRS) can rediscover or outperform human-designed algorithms across cloud scheduling, MoE expert load balancing, LLM-SQL optimization,
Melissa Pan
@melissapan
Apr 12, 2025
Running “Creative Coding” workshops for grade5-8 girls at EYH today - co-design the course with @Ponnapalli95
Melissa Pan
@melissapan
Sep 29, 2025
Replying to @melissapan
blog post link:
melissa-pan.github.io
CS PhD
This season, several friends and family members are applying to PhD programs in Computer Science, thus I have been discussing with people about this topic quite a lot. As, I went through the PhD...
Melissa Pan
@melissapan
Nov 1, 2025
The Sky’s Fun Committee, representing the ppl of sky, just dropped the new lab theme: ⚫️💖 Black Pink x Halloween 🎃🦇 We have: - Gru & the minions - kpop ??? 🫰😉
Melissa Pan
@melissapan
Oct 23, 2025
We are launching a new ADRS blog series to showcase how AI can help systems research 🙌 First up: MoE load balancing ⚖️ AI found algorithmic + engineering optimization! Check out the details in the blog. This case study is led and done by @abmfy_. Bowen is a great undergrad
AI-Driven Research for Systems
@ai4research_ucb
Oct 23, 2025
🚀 We used AI to discover a new algorithm for LLM inference, achieving a 5.0x speedup in MoE load balancing over expert-written code. ✍️ Read the details in our blog post: adrs-ucb.notion.site/moe-load-balan… 📄 Full paper: arxiv.org/abs/2510.06189 💻 Code: github.com/UCB-ADRS/ADRS
Melissa Pan
@melissapan
Oct 11, 2025
Yes! Many cool ways to use LMs in (systems) research, both as a component in the systems or guide the research process itself. Although, I'd question the feasibility / timeline of * fully automated design *🤔 As systems research is way more than just performance optimizations.
Nando de Freitas
@NandoDF
Oct 11, 2025
Optimisation with LLMs is super cool. The crucial thing standing between now and fully automated design and training of powerful LLMs is: Compute (scaling). We’ve already shown that LLMs can do experimental design. The LLMs can search for data, curate it, run ablations,