Sky Computing
Towards Utility Computing for the Cloud
News
November 18, 2025
vLLM is the top open source project on GitHub for 2025
2025’s top projects split between AI infrastructure (vllm, ollama, huggingface/transformers) and enduring ecosystems (vscode, godot, home-assistant).
October 21, 2025
EECS students drive AI innovation as Amazon PhD Fellows
Today, Amazon announced its new AI PhD Fellowship program, offering two years of funding to over 100 PhD students across nine universities. Ten of these inaugural fellowships have been awarded to graduate students from UC Berkeley EECS’ Sky Computing Lab, supporting cutting-edge research in core AI disciplines like machine learning, computer vision, and natural-language processing, ultimately driving innovations essential for the next evolution of practical AI.
October 21, 2025
Amazon launches $68 million AI PhD Fellowship program
“We are thrilled to partner with Amazon to advance open research in AI,” said Joseph E. Gonzalez, a professor of electrical engineering and computer science at UC Berkeley and co-director of the university’s Sky Computing Lab. “Through this fellowship, Amazon and UC Berkeley are investing in the next generation of researchers, and I am excited to see how our PhD students will shape the future of artificial intelligence.”
October 20, 2025
Barbarians at The Gate: How AI is Upending Systems Research
AI is no longer just tuning systems as a “black box.” It’s now rewriting their core algorithms by treating the system as a “white box” and discovering solutions that can outperform human experts in a few hours. This new approach, which we term AI-Driven Research for Systems (ADRS), can automate some of the most tedious parts of research.
October 9, 2025
Lightspeed + Sky Lab: Supporting UC Berkeley’s Next Generation of Entrepreneurs
For decades, UC Berkeley has played a critical role in the technology and startup ecosystem. The trend has only accelerated in the AI era, with alumni having founded some of the most consequential startups of this generation, including OpenAI, Databricks, Perplexity, Skild AI, Reflection, and Physical Intelligence. In fact, according to the latest data from PitchBook, the university’s undergrad alumni have launched more VC-backed startups than any other campus globally. Against this backdrop, and amid National Science Foundation (NSF) changes, Lightspeed and Sky Lab are launching the first formal venture capital partnership in Sky Computing Lab’s history. Lightspeed has made $500k founding contribution to the partnership, whose goal is to sustain pivotal research and translate it into market-ready products and services that address enterprise needs.
Events
December 12, 2025
Sky Seminar: Srinivasan Seshadri (Aerospike) – Architecture of Aerospike: Fast, Scalable, Geo-Replicated, Multi Model Database Supporting Strict Serializable Transactions
Aerospike is a fast, scalable database that supports multiple data models such as a key value store with complex objects, graphs and vectors. Aerospike supports very low latency and high throughput through its novel Hybrid Memory Architecture (HMA). Aerospike supports synchronous replication within …
December 5, 2025
Sky Seminar: Pedro Fonseca (Purdue) – Building reliable and efficient kernels for the next generation workloads
The ever-growing demand for massive data processing and increasing application complexity has resulted in a push towards modular, fine-grained programming and deployment paradigms, such as microservices and serverless computing. These paradigms defy the core design principles of traditional operatin…
November 21, 2025
Sky Seminar: Sujith Ravi (Oracle) – Scaling the AI Frontier at Planet-scale: From Compute and Data to Models, Agents and Multimodal GenAI
…
November 7, 2025
Sky Seminar: Martin Maas (Google DeepMind) – Machine Learning for Systems in the Age of AI
Machine learning and artificial intelligence have the potential to significantly improve computer systems. While research in this area has shown great promise, using modern ML techniques in this context is arguably still in its infancy, and real-world deployments are relatively rare. In this talk, I…
Publications
November 2025
EDIT-Bench: Evaluating LLM Abilities to Perform Real-World Instructed Code Edits.
November 2025
Continuum: Efficient and Robust Multi-Turn LLM Agent Scheduling with KV Cache Time-to-Live.
October 2025
GRAID: Enhancing Spatial Reasoning of VLMs Through High-Fidelity Data Generation.
October 2025
RLBoost: Harvesting Preemptible Resources for Cost-Efficient Reinforcement Learning on LLMs.
October 2025
Efficient Long-context Language Model Training by Core Attention Disaggregation.
October 2025
Budget-aware Test-time Scaling via Discriminative Verification.
October 2025
Reliable Fine-Grained Evaluation of Natural Language Math Proofs.
October 2025
Are Large Reasoning Models Interruptible?
October 2025
Jenga: Effective Memory Management for Serving LLM with Heterogeneity.
October 2025
Pesto: Cooking up High Performance BFT Queries.
October 2025
PrefillOnly: An Inference Engine for Prefill-only Workloads in Large Language Model Applications.
October 2025
Barbarians at the Gate: How AI is Upending Systems Research.
Recent Projects
AgentThink
Search Arena
Sky Computing Story
Berkeley’s computer science division has an ongoing tradition of 5-year collaborative research labs. Recent labs included the AMPLab (ended in 2016) and the RISELab. These labs have had significant impact in both academia and industry. Past labs publish their research at top conferences in systems, databases, and machine learning. On the industrial side, AMPLab and RISELab fostered several successful startups (Databricks, Opaque, Ponder, Anyscale, to name a few). We are excited to announce the Berkeley Sky Computing Lab where we will strike to make cloud computing a true commodity.
Context
The Sky Computing Lab represents the next chapter of data-intensive systems research at Berkeley. Recent years have seen the explosion of cloud computing. Applications are moving their data and computation to the cloud; on-premise services are dying. In doing so, companies have to make difficult choices between the myriad of cloud providers, each with different services or hardware. Lock-in, whether through artificial migration costs, legal constraints or engineering baggage is real. In the Sky Computing Lab, we will leverage distributed systems, programming languages, security, and machine learning to decouple the services that a company wants to implement from the choice of a specific cloud. Much like the Internet today, cloud computing should be an undifferentiated commodity. Applications should run seamlessly on any or multiple clouds.
Mission
Our mission in the Sky Computing Lab is to transform the cloud into an undifferentiated commodity and ease application burden. As in previous labs, we’re all in — working on everything from basic research to software development, all in the Berkeley tradition of open publication and open source software. Our founding team consists of experts in distributed systems, machine learning, security and programming languages. We’ll use this space to lay out our ideas and progress as we go.
Commitment to Diversity
Sky Computing is guided by Berkeley’s Principles of Community and is committed to providing a safe and caring research environment for every member of our community. We believe that a diverse student body, faculty, and staff are essential to the open exchange of ideas that Sky Computing Lab is founded on.
Our head is in the cloud. We are heading for the SKY.
Sponsors
