Publications

I’m working on publishing technical papers and giving talks on storage systems being built for the new wave of AI/ML infrastructure.

Areas Currently Working On

AI Storage Optimization

• RDMA performance in production: How to get sub-microsecond latency for GPU training workloads with focus on Object Store
• Tier-0 caching: Pre-fetching and eviction policies for transformer training access patterns
• Multi-protocol storage: Making NFS and S3 work on the same data without breaking consistency

Distributed Systems Performance

• Storage benchmarking: Building tools that measure what actually matters for ML workloads (not just random IOPS)
• High-throughput streaming: Scaling data pipelines to billions of events using Akka Streams and reactive patterns

Looking For

• Co-authors, Speaking opportunities, Peer review on distributed storage systems , ML infrastructure and Benchmarking etc at conferences like at FAST, OSDI, SoCC, MLSys etc.
• Open to Collaboration on open-source benchmarking tools or storage projects

If you’re organizing conferences, running workshops, or writing papers in this space or if you just want to talk about storage performance, feel free to reach out.

Academic & Industry Interests

I have been trying to learn and draw inspiration from foundational distributed systems papers & work like Dynamo, Spanner, GFS, Raft, Tao, Memcached etc and aim to bridge academic research with real-world production challenges in AI infrastructure.

Interested in collaboration, co-authorship, speaking opportunities or if you are organizing conferences/workshops in these areas? Reach out.