-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
executable file
·330 lines (308 loc) · 19.6 KB
/
index.html
File metadata and controls
executable file
·330 lines (308 loc) · 19.6 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
<!DOCTYPE HTML>
<html lang="en"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Michael Luo</title>
<meta name="author" content="Michael Luo">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
<link rel="icon" type="image/png" href="images/seal_icon.png">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.1/css/all.min.css">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<style>
.icon-link { font-size: 22px; margin: 0 4px; color: #1772d0; text-decoration: none; }
.icon-link:hover { color: #f09228; text-decoration: none; }
.topnav { background-color: #f5f5f5; border-bottom: 1px solid #e0e0e0; padding: 0; margin: 0; text-align: right; position: sticky; top: 0; z-index: 100; }
.topnav a { display: inline-block; padding: 12px 20px; color: #555; font-size: 14px; font-family: 'Lato', Verdana, Helvetica, sans-serif; text-decoration: none; font-weight: 700; letter-spacing: 0.5px; }
.topnav a:hover { color: #1772d0; }
.topnav a.active { color: #1772d0; border-bottom: 2px solid #1772d0; }
</style>
</head>
<body>
<div class="topnav">
<a href="index.html" class="active">Home</a>
<a href="publications.html">Publications</a>
<a href="blog.html">Blog</a>
<a href="data/Michael_Luo_CV_PhD.pdf">CV</a>
</div>
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:0px">
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:2.5%;width:63%;vertical-align:middle">
<name>Michael Luo</name>
<p>I am currently building a stealth startup. My interests lie broadly in building next generation AI systems and models (e.g. efficient serving and post-training). Previously, I was at <a href="https://deepmind.google/">Google DeepMind</a>, where I worked on efficient agentic systems for serving and post-training.
</p>
<p>I created the <a href="https://www.agentica-project.com">Agentica Project</a>, an open-source initiative for post-training language agents via RL, producing models like <a href="https://pretty-radio-b75.notion.site/DeepScaleR-Surpassing-O1-Preview-with-a-1-5B-Model-by-Scaling-RL-19681902c1468005bed8ca303013a4e2">DeepScaleR</a>, <a href="https://pretty-radio-b75.notion.site/DeepCoder-A-Fully-Open-Source-14B-Coder-at-O3-mini-Level-1cf81902c14680b3bee5eb349a512a51">DeepCoder</a>, and <a href="https://pretty-radio-b75.notion.site/DeepSWE-Training-a-Fully-Open-sourced-State-of-the-Art-by-Scaling-RL-22281902c1468193aabbe9a8c59bbe33">DeepSWE</a> with over 2M+ downloads. Our training recipes are published on the <a href="https://github.com/agentica-project/rllm">rLLM</a> framework.
</p>
<p>I received my <b>PhD</b> in EECS from <a href="https://eecs.berkeley.edu/">UC Berkeley</a>, where I was advised by <a href="https://people.eecs.berkeley.edu/~istoica/">Ion Stoica</a> and was part of <a href="https://sky.cs.berkeley.edu/">SkyLab</a> and <a href="https://bair.berkeley.edu/">BAIR</a>. I also hold an M.S. and B.S. (CS & Business) from UC Berkeley.
</p>
<p style="text-align:center">
<a href="mailto:michael.luo@berkeley.edu" class="icon-link" title="Email"><i class="fas fa-envelope"></i></a>
<a href="https://scholar.google.com/citations?hl=en&user=XpO6-kEAAAAJ" class="icon-link" title="Google Scholar"><i class="ai ai-google-scholar"></i></a>
<a href="https://github.com/michaelzhiluo/" class="icon-link" title="GitHub"><i class="fab fa-github"></i></a>
<a href="https://www.linkedin.com/in/michaelzhiluo/" class="icon-link" title="LinkedIn"><i class="fab fa-linkedin"></i></a>
<a href="https://x.com/michaelzluo" class="icon-link" title="X / Twitter"><i class="fab fa-x-twitter"></i></a>
</p>
</td>
<td style="padding:2.5%;width:40%;max-width:40%">
<a href="images/michael_luo_1.jpg"><img style="width:100%;max-width:100%" alt="profile photo" src="images/michael_luo_1.jpg" class="hoverZoomLink"></a>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px 20px 5px 20px;width:100%;vertical-align:middle">
<heading>Selected Publications</heading>
<p style="margin-bottom:0;">
For a full list, see my <a href="publications.html">publications page</a> or <a href="https://scholar.google.com/citations?hl=en&user=XpO6-kEAAAAJ">Google Scholar</a>.
</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:10px 20px;width:25%;vertical-align:middle">
<img src="images/agentica_logo.png" alt="Agentica Project" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:10px 20px;width:75%;vertical-align:middle">
<papertitle><a href="https://www.agentica-project.com">Agentica Project</a></papertitle>
<p>
An open-source initiative for post-training language agents via reinforcement learning. Agentica models have achieved over <strong>2M+ downloads</strong> on HuggingFace. We've optimized RL systems and published training recipes on the <a href="https://github.com/agentica-project/rllm">rLLM</a> project, now maintained by academia, with over <strong>5K+ stars</strong>.
<br>
<span style="display:inline-block;margin-top:2px;line-height:22px;">
<a href="https://www.agentica-project.com" class="icon-link" title="Website"><i class="fas fa-globe" style="vertical-align:middle;"></i></a>
<a href="https://huggingface.co/agentica-org" class="icon-link" title="HuggingFace" style="vertical-align:middle;"><img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" style="width:20px;height:20px;vertical-align:middle;position:relative;top:-2px;"></a>
<a href="https://github.com/agentica-project/rllm" class="icon-link" title="GitHub"><i class="fab fa-github" style="vertical-align:middle;"></i></a>
<a href="https://x.com/Agentica_" class="icon-link" title="X / Twitter"><i class="fab fa-x-twitter" style="vertical-align:middle;"></i></a>
</span>
</p>
</td>
</tr>
<tr>
<td style="padding:6px 8px 6px 60px;width:15%;vertical-align:middle">
<img src="images/deepscaler.png" alt="DeepScaleR" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:6px 20px;vertical-align:middle">
<a href="https://pretty-radio-b75.notion.site/DeepScaleR-Surpassing-O1-Preview-with-a-1-5B-Model-by-Scaling-RL-19681902c1468005bed8ca303013a4e2">
<papertitle>DeepScaleR: Surpassing o1-preview with a 1.5B Model by Scaling RL</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Sijun Tan, Justin Wong, Xiaoxiang Shi, William Y. Tang, Manan Roongta, Colin Cai, Jeffrey Luo, Li Erran Li, Raluca Ada Popa, Ion Stoica
<br>
<em>Notion Blog, Feb. 2025</em>
<a href="https://pretty-radio-b75.notion.site/DeepScaleR-Surpassing-O1-Preview-with-a-1-5B-Model-by-Scaling-RL-19681902c1468005bed8ca303013a4e2">Blog</a>
</td>
</tr>
<tr>
<td style="padding:6px 8px 6px 60px;width:15%;vertical-align:middle">
<img src="images/deepcoder.png" alt="DeepCoder" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:6px 20px;vertical-align:middle">
<a href="https://pretty-radio-b75.notion.site/DeepCoder-A-Fully-Open-Source-14B-Coder-at-O3-mini-Level-1cf81902c14680b3bee5eb349a512a51">
<papertitle>DeepCoder: A Fully Open-Source 14B Coder at o3-mini Level</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Sijun Tan, Roy Huang, Ameen Patel, Alpay Ariyak, Qingyang Wu, Xiaoxiang Shi, Rachel Xin, Colin Cai, et al.
<br>
<em>Notion Blog, Apr. 2025</em>
<a href="https://pretty-radio-b75.notion.site/DeepCoder-A-Fully-Open-Source-14B-Coder-at-O3-mini-Level-1cf81902c14680b3bee5eb349a512a51">Blog</a>
</td>
</tr>
<tr>
<td style="padding:6px 8px 6px 60px;width:15%;vertical-align:middle">
<img src="images/deepswe.png" alt="DeepSWE" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:6px 20px;vertical-align:middle">
<a href="https://pretty-radio-b75.notion.site/DeepSWE-Training-a-Fully-Open-sourced-State-of-the-Art-by-Scaling-RL-22281902c1468193aabbe9a8c59bbe33">
<papertitle>DeepSWE: Training a State-of-the-Art Coding Agent from Scratch by Scaling RL</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Naman Jain, Jaskirat Singh, Sijun Tan, Ameen Patel, Qingyang Wu, Alpay Ariyak, Colin Cai, Tarun Venkat, et al.
<br>
<em>Notion Blog, Jun. 2025</em>
<a href="https://pretty-radio-b75.notion.site/DeepSWE-Training-a-Fully-Open-sourced-State-of-the-Art-by-Scaling-RL-22281902c1468193aabbe9a8c59bbe33">Blog</a>
</td>
</tr>
<tr>
<td style="padding:6px 8px 6px 60px;width:15%;vertical-align:middle">
<img src="images/rllm.png" alt="rLLM" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:6px 20px;vertical-align:middle">
<a href="https://pretty-radio-b75.notion.site/rLLM-A-Framework-for-Post-Training-Language-Agents-21b81902c146819db63cd98a54ba5f31">
<papertitle>rLLM: A Framework for Post-Training Language Agents</papertitle>
</a>
<br>
Sijun Tan*, <strong>Michael Luo*</strong>, Colin Cai*, Tarun Venkat, Kyle Montgomery, Aaron Hao, Tianhao Wu, Arnav Balyan, Manan Roongta, Chenguang Wang, Li Erran Li, Raluca Ada Popa, Ion Stoica
<br>
<em>Notion Blog, Jun. 2025</em>
<a href="https://pretty-radio-b75.notion.site/rLLM-A-Framework-for-Post-Training-Language-Agents-21b81902c146819db63cd98a54ba5f31">Blog</a> |
<a href="https://github.com/agentica-project/rllm">Code</a>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/autellix.png" alt="Autellix" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://www.usenix.org/conference/nsdi26/presentation/luo">
<papertitle>Autellix: An Efficient Serving Engine for LLM Agents as General Programs</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Xiaoxiang Shi, Colin Cai, Tianjun Zhang, Justin Wong, Yichuan Wang, Chi Wang, Yanping Huang, Zhifeng Chen, Joseph E. Gonzalez, Ion Stoica
<br>
<em>USENIX Symposium on Networked Systems Design and Implementation (<strong>NSDI</strong>), 2026</em>
<br>
<a href="https://arxiv.org/abs/2502.13965">Arxiv</a> |
<a href="https://www.usenix.org/conference/nsdi26/presentation/luo">Paper</a>
<p></p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/stylus.gif" alt="Stylus" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2404.18928">
<papertitle>Stylus: Automatic Adapter Selection for Diffusion Models</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Justin Wong, Brandon Trabucco, Yanping Huang, Joseph E. Gonzalez, Zhifeng Chen, Ruslan Salakhutdinov, Ion Stoica
<br>
<em>Neural Information Processing Systems (<strong>NeurIPS</strong>), 2024</em> <span style="color:#d14836;font-weight:bold;">Oral</span>
<br>
<a href="https://arxiv.org/abs/2404.18928">Arxiv</a> |
<a href="https://neurips.cc/virtual/2024/oral/98000">Talk</a>
<p></p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/starburst.png" alt="Starburst" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://www.usenix.org/conference/atc24/presentation/luo">
<papertitle>Starburst: A Cost-aware Scheduler for Hybrid Cloud</papertitle>
</a>
<br>
<strong>Michael Luo</strong>, Siyuan Zhuang, Suryaprakash Vengadesan, Romil Bhardwaj, Justin Chang, Eric J. Friedman, Scott Shenker, Ion Stoica
<br>
<em>USENIX Annual Technical Conference (<strong>USENIX ATC</strong>), 2024</em> <span style="color:#d14836;font-weight:bold;">Best Paper Award</span>
<br>
<a href="https://www.usenix.org/conference/atc24/presentation/luo">Paper</a>
<p></p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/skypilot.png" alt="SkyPilot" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://www.usenix.org/conference/nsdi23/presentation/yang-zongheng">
<papertitle>SkyPilot: An Intercloud Broker for Sky Computing</papertitle>
</a>
<br>
Zongheng Yang, Zhanghao Wu, <strong>Michael Luo</strong>, Wei-Lin Chiang, Romil Bhardwaj, Woosuk Kwon, Siyuan Zhuang, Frank Sifei Luan, Gautam Mittal, Scott Shenker, Ion Stoica
<br>
<em>USENIX Symposium on Networked Systems Design and Implementation (<strong>NSDI</strong>), 2023</em>
<br>
<a href="https://www.usenix.org/conference/nsdi23/presentation/yang-zongheng">Paper</a> |
<a href="https://github.com/skypilot-org/skypilot">Code</a>
<p></p>
</td>
</tr>
</tbody></table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20"><tbody>
<tr>
<td>
<heading>Education</heading>
</td>
</tr>
</tbody></table>
<table width="100%" align="center" border="0" cellpadding="20"><tbody>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img style="width:105%;max-width:105%" src="images/berkeley_eecs.png">
</td>
<td width="75%" valign="center">
<p>
<b><a href="https://eecs.berkeley.edu/">University of California, Berkeley</a></b> <br>
Ph.D. in Electrical Engineering and Computer Science, Aug 2021 - Fall 2025
<br>
<span style="color:#777;">Advised by <a href="https://people.eecs.berkeley.edu/~istoica/">Ion Stoica</a></span>
</p>
<p>
<b><a href="https://eecs.berkeley.edu/">University of California, Berkeley</a></b> <br>
M.S. in Electrical Engineering and Computer Science, Aug 2020 - May 2021
<br>
<span style="color:#777;">Advised by <a href="https://people.eecs.berkeley.edu/~istoica/">Ion Stoica</a> and <a href="https://goldberg.berkeley.edu/">Ken Goldberg</a></span>
</p>
<p>
<b><a href="https://eecs.berkeley.edu/">University of California, Berkeley</a></b> <br>
B.S. in EECS and Business Administration (<em>Summa Cum Laude</em>), Aug 2016 - May 2020
<br>
<span style="color:#777;">GPA: 3.98/4</span>
</p>
</td>
</tr>
</tbody></table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20"><tbody>
<tr>
<td>
<heading>Work Experience</heading>
</td>
</tr>
</tbody></table>
<table width="100%" align="center" border="0" cellpadding="20"><tbody>
<tr>
<td style="padding:12px 20px;width:15%;vertical-align:middle"><img style="width:100%;max-width:80px;" src="images/deepmind.png"></td>
<td width="75%" valign="center">
<b><a href="https://deepmind.google/">Google DeepMind</a></b>
<br>
Research Scientist, Sept 2023 - March 2025
<br>
<span class="gs_gray" style="color:#777;">SaxML Team; Developed and researched efficient agentic systems for serving and post-training.</span>
</td>
</tr>
<tr>
<td style="padding:12px 20px;width:15%;vertical-align:middle"><img style="width:100%;max-width:80px;" src="images/anyscale.png"></td>
<td width="75%" valign="center">
<b><a href="https://www.anyscale.com/">Anyscale</a></b>, Ray Core Team
<br>
Software Development Engineer Intern, June 2020 - Jan 2021
<br>
<span style="color:#777;">Scaled distributed RL training and developed asynchronous RL algorithms with Ray/RLlib.</span>
</td>
</tr>
<tr>
<td style="padding:12px 20px;width:15%;vertical-align:middle"><img style="width:100%;max-width:80px;" src="images/amazon.jpg"></td>
<td width="75%" valign="center">
<b><a href="https://www.amazon.com/">Amazon AI</a></b>
<br>
Machine Learning Engineer Intern, June 2019 - Aug 2019
<br>
<span style="color:#777;">Created the first embedding/RAG based recommendation system for Amazon Ads.</span>
</td>
</tr>
<tr>
<td style="padding:12px 20px;width:15%;vertical-align:middle"><img style="width:100%;max-width:80px;" src="images/meraki.png"></td>
<td width="75%" valign="center">
<b><a href="https://meraki.cisco.com/">Cisco Meraki</a></b>, Smart Camera Team
<br>
Computer Vision Engineer Intern, June 2018 - Aug 2018
<br>
<span style="color:#777;">Developed model-based object detection to track individuals across Meraki cameras.</span>
</td>
</tr>
</tbody></table>
<br>
</td>
</tr>
</table>
</body>
</html>