About
Building verifiable, frictionless, and high-performance abstractions for…
Experience
Education
-
UC San Diego
-
Completed advanced coursework alongside research in distributed systems and architectures within Prof. Yiying Zhang's WukLab (wuklab.io). Also served as Teaching Assistant for the following courses:
• CSE 124: Networked Systems (Fall 2021)
• CSE 224: Graduate Networked Systems (Winter 2022)
• CSE 260: Parallel Computation (Spring 2022) -
-
-
-
-
-
Volunteer Experience
-
Org Admin & Mentor for P4 GSoC
Google Summer of Code
- Present 2 years 8 months
Science and Technology
Together with Fabian Ruffy, Steffen Smolka, and Hari Thantry, we founded P4 org's participation in Google Summer of Code (GSoC). I have served as one of the org admins since 2024, and a mentor for several projects over the years.
Other similar profiles
Explore more posts
-
Petar Georgiev
Can we draw an analogy between the compute–memory tradeoff and the inference–representation relationship? If you get stuck in your reasoning, try changing the coordinate system or exploring how a model’s inference power might be enhanced with more suitable representations. Learn more about complexity, intelligence, and emergence from Prof. David Krakauer. #ai #compute #complexity https://lnkd.in/dmq7Cjn4
-
Turing
In our latest Turing Podcast episode, Lilin Wang, Engineering Director at Turing, explores how SWE Bench is redefining what it means to evaluate large language models. 🎙️ 🤖 This benchmark moves beyond syntax checks to test a model’s ability to reason through real-world engineering tasks like bug fixes and feature development. Lilin shares how Turing is using data expansion and trajectory data to improve model performance and why SWE Bench presents deeper challenges than benchmarks like Human Eval. We also look ahead to the future of AI in software engineering: one where models act as junior engineers, and humans supervise high-level design and planning. Listen in as we map the path toward O5-level reasoning capabilities. 🎧 Apple podcast: https://lnkd.in/dKbpKmSC 🎧 Spotify: https://lnkd.in/dkdmkm2R #TuringPodcast #SWEBench #LLMs #FutureOfAI #TechPodcast
2 Comments -
Dan Zhang
Want to know how Google DeepMind scales Gemini inference? Rahul Arya, ranked #82 on the Metis List, is giving an invited talk at the ML for Systems workshop at NeurIPS! He will be discussing Trillium, Ironwood, and trade-offs to achieve new speeds in LLM serving. https://mlforsystems.org/ Title: Advances in LLM Serving Efficiency at Scale Bio: Rahul Arya is a research engineer at Google DeepMind contributing to the training and inference performance of Gemini models. He previously worked on the XLA:TPU compiler.
3 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content