I am a Ph.D. student in Computer Science at UC Berkeley where I am advised by Joseph Gonzalez, Prabal Dutta, and Ion Stoica. I am broadly interested in ML-Systems, especially around LLM post-training, and tool use for LLMs. I am affiliated with the Sky Computing Lab (previously RISE), Lab11, and Berkeley AI Research (BAIR).
I created and lead the Gorilla, GoEx, RAFT, OpenFunctions, and Berkeley Function Calling Leaderboard (BFCL) projects. Learn more about Gorilla: [Website] [Github] [Twitter]
My dissertation [talk] on the evolution of Agentic LLMs from function calling to truly autonomous agents.
News:
[Mar 2018] Presented our work to Bill Gates![Dec 2024] Gorilla published at NeurIPS 2024
[Nov 2024] LLoCO published at EMNLP 2024
[Oct 2024] Launched Agent Arena by 🦍 Gorilla X LMSYS Chatbot Arena. Check it out
[Sep 2024] Released BFCL v3 with multi-turn. Release blog here
[Aug 2024] Released BFCL v2 with community contributed APIs. Release blog
[Apr 2024] Microsoft covers GoEx in their AI platform blog!
[Apr 2024] Release GoEx: A runtime for Agentic LLMs. Read more in our release blog
[Mar 2024] Meta and Microsoft cover our RAFT paper!
[Mar 2024] Released RAFT: Retrieval-Augmented Fine-Tuning for LLMs.
[Feb 2024] Released Berkeley Function Calling Leaderboard (BFCL). Read more in our release blog
[Feb 2024] Released Gorilla OpenFunctions-v2. New SOTA in function calling.
[Dec 2023] Released Gorilla OpenFunctions. Check out our release blog
[Nov 2023] Presenting Gorilla at at UCL NLP, Meta, and DeepMind London!
[Oct 2023] Attending the Dagstuhl seminar on EdgeAI!
[May 2023] Released Gorilla LLM!
[Apr 2023] Talk on POET at Harvard Systems + Theory group
[Apr 2023] Skyplane published at NSDI 2023!
[Dec 2022] Talk at Microsoft Research India, and Google Research, Bangalore
[Sep 2022] IEEE Spectrum article on POET!
[Sep 2022] Led the Skyplane tutorial for Skycamp 2022
[Sep 2022] Skyplane accepted to NSDI 2023
[Sep 2022] I will be talking about ML on Edge at Princeton and CMU
[Aug 2022] Presented POET at Google Federated Learning Talks, and Google Language Seminar
[Aug 2022] Presented Galaxy and On-device ML at the Conix workshop at UW, Seattle
[Jul 2022] Presented POET as spotlight at ICML 2022! Camera ready on arXiv
[May 2022] Presented POET and Skyplane posters at RISE Retreat [Tahoe, CA]
[May 2022] POET accepted to ICML '22! Camera ready coming soon..
[May 2022] I will be interning with the Brain and Cloud teams at Google this Summer
[Apr 2022] Where the Sidewalk Ends: Privacy of Opportunistic Backhaul presented at EuroSec'22
[May 2021] I will be interning with the Core-ML team at Amazon Science this Summer
[Dec 2020] Embeddings for Indoor Navigation presented at NeurIPS'20 ORLR Workshop
[Jan 2020] Attending The Quantum Wave in Computing Boot Camp at Simons Institute
[Jan 2020] Presented poster at RISE Retreat [Monterey, CA]
[Jan 2020] Gave a talk at VMare Retreat [Palo Alto, CA]
[Jul 2019] GesturePod accepted to UIST 2019!
[Dec 2018] ZD Net covers our work at (NeurIPS) NIPS 2018
[Nov 2018] We will be presenting our work at (NeurIPS) NIPS 2018!
[Nov 2018] Demonstrated programmable gesture recognition on Xbox controllers with EdgeML
[Oct 2018] GesturePod implementation and simulation OSS
[Dec 2017] Our work covered by Financial Express and Microsoft AI blog.
Here are some key projects I have co-created:
Star
Gorilla
: Large Language Model Connected with Massive APIs
Star
POET
: Neural Network training on edge devices
Star
EdgeML
: Neural Network inference on edge devices
Star
Gorilla-CLI
: LLMs for your Command Line
Few Media Highlights of my work
- Microsoft and Meta blog on our RAFT paper
- The Daily Californian newspaper features Gorilla
- Gorilla paper walkthrough, podcast, and How Weviate trained their own Gorilla
- My advisor interviewed me: Generating Conversations with Prof. Joseph Gonzalez
- Gorilla on Youtube
- IEEE Spectrum article on POET
- Adafruit blog on Paul's Workshop at The Things Conference, UK
- Arduino and Hackster at ACM UIST 2019
- ZD Net at NeurIPS 2018
- Satya Nadella in an interview with Indian media [5:45 - 7:30]
- Financial Express
Talks
- “Gorilla and Connecting Large Language Models to the outside world”
- Google Research (Bengaluru, India), December 2023
- Microsoft Research India (Bengaluru, India), December 2023
- Deepmind (London, United Kingdom), November 2023
- UCL NLP Seminar Series (London, United Kingdom), November 2023
- Meta (London, United Kingdom), November 2023
- Sky Camp (Berkeley, CA), October 2023
- Ray Summit (San Francisco, CA), September 2023
- Kong API Summit (San Francisco, CA), September 2023
- Simon’s Institute for Theory of Computing (Berkeley, CA), August 2023
- Llamaindex Webinar (San Francisco, CA), August 2023
- Mosaic ML (San Francisco, CA), August 2023
- Intel (Santa Clara, CA), August 2023
- Microsoft Research Redmond (Redmond, WA), June 2023
- Berkeley LLM Hackathon (Berkeley, CA), June 2023
- Apple (Cupertino, CA), May 2023
- Sky Summer Retreat (Tahoe, CA), May 2023
- “POET: Training Neural Networks for the Bleeding Edge!”
- Dagstuhl Seminar (Wadern, Germany), October 2023
- Harvard Systems + Theory group (Boston, MA), April 2023
- Sky Winter Retreat (Monterey, CA), January 2023
- Google Research (Bengaluru, India), December 2022
- Microsoft Research (Bengaluru, India), December 2022
- Carnegie Mellon University (Pittsburgh, PA), September 2022
- Princeton University (Princeton, NJ), September 2022
- Google Federated Learning Talks (Mountain View, CA), August 2022
- University of Washington (Seattle, WA), July 2022
- “GesturePod and on device ML”
- UIST 2019 (New Orleans, LA, USA), October 2019
- Microsoft Research Techfest (Redmond, WA), March 2019
- VMare Research 2019 (Monterey, CA), January 2019
- RISE Retreat 2019 (Monterey, CA), January 2019
- Microsoft Research (Bengaluru, India), January 2019
Miscellaneous: I race go-karts all year-round, ski in winters, and sail in summers. If we are going sailing, please read this.