1. X
  2. Johnson Shi
Log inSign up
Johnson Shi
29 posts
user avatar
Johnson Shi
@johnsonshi86
Senior Product Manager @Microsoft @Azure ⚡️👨‍💻 Scaling up AI Infrastructure on Azure 🌥️
Seattle, WA
johnsonshi.com
Joined June 2017
4,307
Following
895
Followers
RepliesRepliesArticlesArticlesMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • user avatar
    Johnson Shi
    @johnsonshi86
    Jul 22
    Yep had the same reaction to this paper. There’s elegance and cost reduction in separating concerns that are often bundled together (durable storage, sequencing/consensus, state materialization). I wanted to make those ideas more concrete for engineers building in the Azure/.NET
    user avatar
    Gwen (Chen) Shapira
    @gwenshap
    Jul 19
    Always great to see papers that are directly relevant to my interests and work in OSDI (in this case, using S3 as storage layer effectively). Even better when I know and respect most of the authors! And the best part is that unlike many industry papers, this one is focused and
    337
  • user avatar
    Johnson Shi
    @johnsonshi86
    Jul 8
    Kubernetes has quietly become the substrate for AI, yet those AI workloads depend on a fragile step: failed or throttled pulls leading to GPU cluster underutilization. @Azure Chaos Studio (now public preview) makes it easy to inject failures against @azurekubernetes clusters.
    user avatar
    Johnson Shi
    @johnsonshi86
    Jul 8
    Article cover image
    Article
    Validating AKS to Registry Resilience with Azure Chaos Studio
    Kubernetes has quietly become a leading substrate for AI inferencing and harness engineering. As more AI workloads move onto AKS for Azure customers, resilience stops being a nice to have and becomes...
    449
  • user avatar
    Johnson Shi
    @johnsonshi86
    Jul 8
    Article cover image
    Article
    Validating AKS to Registry Resilience with Azure Chaos Studio
    Kubernetes has quietly become a leading substrate for AI inferencing and harness engineering. As more AI workloads move onto AKS for Azure customers, resilience stops being a nice to have and becomes...
    528
  • user avatar
    Johnson Shi
    @johnsonshi86
    Jul 1
    Came across DR-DCI via @jobergum's talk at @aiDotEngineer. It's tackling a problem that comes up constantly once you're building agents: how does an agent actually search over a large and distributed set of documents. Traditional RAG has known limits. It chunks documents, embeds
    user avatar
    Johnson Shi
    @johnsonshi86
    Jul 1
    Article cover image
    Article
    DR-DCI: Fast Corpus Retrieval as a Harness Engineering Differentiator for Inference Providers
    DR-DCI is an optimization built on top of RAG for letting agents run precise, verifiable search across large document collections without scanning the whole corpus on every query. It keeps a RAG-style...
    9.1K
  • user avatar
    Johnson Shi
    @johnsonshi86
    Jul 1
    Article cover image
    Article
    DR-DCI: Fast Corpus Retrieval as a Harness Engineering Differentiator for Inference Providers
    DR-DCI is an optimization built on top of RAG for letting agents run precise, verifiable search across large document collections without scanning the whole corpus on every query. It keeps a RAG-style...
    5.9K