Kenneth Li

Software Engineer

San Francisco, California, United States10 yrs 1 mo experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in high-performance distributed systems.
Led successful projects improving throughput by 2x.
Proficient in Rust and C++ for system-level programming.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer in the SaaS industry.

Contact

Skills

Core Skills

Distributed SystemsInformation Retrieval

Other Skills

Rust (Programming Language)C++Data IngestionGraph DatabasesPlatform DevelopmentRecommender SystemsQuery OptimizationSoftware EngineeringWeb ServicesSoftware DevelopmentComputer ScienceWeb DesignWeb DevelopmentJavaHTML5

About

Passionate about high-performance distributed systems in the realms of indexing, retrieval, storage, and databases. Also an avid user of Rust.

Experience

10 yrs 1 mo

Total Experience

2 yrs 6 mos

Average Tenure

6 yrs 11 mos

Current Experience

3 roles

Senior Staff Software Engineer

Promoted

Mar 2025 – Present · 1 yr 2 mos

Building high performance indexing, retrieval, and storage systems that power LinkedIn's core product experiences.
Caspian (current)
Distributed storage layer that provides high availability and durability for LinkedIn's storage workloads (Rust).
We are still in progress of building and migrating initial blob storage use cases.
NemoDB
A low latency, in-memory retrieval engine that powers LinkedIn's Feed (Rust)
It was designed to support complex filtering predicates, graph-like data modeling, and embedding-based retrieval (i.e vector search).
Led the team from prototype to 100% ramp of Feed use cases, achieving a 2x throughput improvement compared to the legacy Feed serving stack
Check out our blog post: https://www.linkedin.com/blog/engineering/infrastructure/fishdb-a-generic-retrieval-engine-for-scaling-linkedins-feed
LIquid
A distributed in-memory graph index that powers LinkedIn's economic graph and serves 1M+ qps and 200B+ edges (C++)
Worked on query optimization and automatic data ingestion of relational tables