Maintainer of MuopDB, a vector database built from scratch in Rust.Repository: https://github.com/hicder/muopdbMuopDB currently supports:* Index type: HNSW, IVF, SPANN. All on-disk with mmap.* Quantization: PQ (product quantization).Architecturally, MuopDB uses aggregator-leaf pattern, and a separate binary to build the index.

Working on Amplitude's in-house OLAP database that powers Amplitude products.

Working on Facebook's next generation of container and package distribution, supporting various use-cases: Host Management and AI Infra.Published in VLDB 2022: Meta's Next-generation Realtime Monitoring and Analytics Platform

Take a break to take care of my baby daughter Catherine

I work on the scalability and reliability of Rockset's platform. My projects span across the entire stack, including our HTTP servers, ingestion pipeline, storage engine, and networking. A few highlights of my work:* RocksDB-Cloud's remote compactionWe separate ingest compute from query compute in order to scale out our storage engine. With this, query is not affected by background ingest activities, and RocksDB-Cloud can elastically adapt to spikes in data ingestion.Blog: https://rockset.com/blog/remote-compactions-in-rocksdb-cloud/Talk: https://youtu.be/XfwxUBL8xT8?t=2569* Investigating and squashing 502/504 errors in our networking stackThis work vastly improves the reliability of our customer-facing SLA, by squashing various random 502s and 504s errors that occur in our system. Sometimes these errors happen before the queries even reach our HTTP servers. This is a 5-month hunt to squash multiple misconfigurations in various different parts of our system, working closely with multiple teams and requiring deep understanding of our entire stack.https://rockset.com/blog/fixing-502-504-errors/

In my time at Facebook, I worked on various teams, including:* Search Infrastructure: https://research.facebook.com/publications/unicorn-a-system-for-searching-the-social-graph/* Realtime Monitoring: https://www.vldb.org/pvldb/vol8/p1816-teller.pdf* Realtime Analytics: http://www.vldb.org/pvldb/vol6/p1057-wiener.pdf* Graph Indexing: https://engineering.fb.com/2016/03/18/data-infrastructure/dragon-a-distributed-graph-query-engine/I mostly focused on scalability and reliability of the systems. My background spans from data infra to online realtime serving.

ODS Monitoring, Data Infrastructure

Data and Fundamentals, OS Core.

Worked to improve reliability of MySQL