Xinye Tao (`陶新野`)

Skills

Programming Language: experienced in C/C++/Rust, familiar with Python/Java/Go/Verilog HDL.
Development Skills: experienced in parallel programming with C++/Rust/Java, GPU programming with OpenGL and CUDA, well familiar with database and distributed system.
Language: Chinese (native), English (fluent, CET-6 629).
Experienced with working remotely and collaborating with remote teams.
Familiar with public cloud platforms, especially the performance characteristics of the storage stack.

Metabit Trading - Data Platform (Software Engineer; Oct, 2023 ~ now)

PingCAP - Storage (Database Developer; Aug, 2018 ~ Oct, 2023)

The most active contributor to TiKV in the three-year timespan.
Tackled the performance stability issue of TiKV. Implemented a unified physical I/O scheduling framework that reduces tail latency by 30% during data analyze and scaling.
Conducted the early adaptation of TiKV for public clouds.
Developed a new log storage engine Raft Engine as a replacement for RocksDB. It reduces I/O bandwidth usage by 30% and tail latency by 20%. On AWS cloud, it improves TiDB transaction throughput by 25%. It has a unit test coverage of 98%. It is also being used by some other commercial databases.
Maintaining and optimizing PingCAP’s own fork of RocksDB. Found and fixed multiple performance issues of upstream RocksDB. Implemented major features that enable TiKV to run thousands of RocksDB instances simultaneously on one server.
Lead the tech sharing program in the team. Hosted sharing sessions at CNCF Maintainer Track and USTC Event. Lead the research collaboration program with USTC.
Optimized CI infrastructure to cut down 80% resource usage and 40% build time.
Received Ownership Award in 2022 (top 10/250). Received FY23 annual high performance recognition (top 2/250).

Microsoft Research Asia - Intelligent Cloud & Edge (Research Intern; Nov, 2019 – Jun, 2020)

Worked on building new-generation database systems with decoupled service components.
Implemented and improved data replication for the Fugue system.
Received a reward of Excellence from the Stars of Tomorrow Internship Program.

Ant Financial – OceanBase System Group (Developer Intern; Jun, 2019 ~ Aug, 2019)

Exploring the performance limits of Oracle and PostgreSQL under TPC-C benchmarks.
Optimized the coroutine component in OceanBase’s internal library.

PingCAP – TiKV Engine Team (Developer Intern; Apr, 2019 ~ Nov, 2019)

Worked on making Titan, a RocksDB plugin for key-value separation, generally available.
Implemented adaptive RPC request batching for TiKV.

Lab for Internet and Security Technology in Zhejiang University (Research Intern; Oct, 2018 ~ Mar, 2019)

Mentored by A/Prof Kai Bu. Worked on a research project aiming at optimizing durable hardware transactional memory on NVM.

State Key Lab of CAD & CG (Developer Intern; Jul, 2018 ~ Sept, 2018)

Mentored by A/Prof Hongzhi Wu. Participated in a research project on reflectance acquisition with deep-learning.
Implemented a close to real-time physically based renderer using CUDA and NVIDIA OptiX.
The renderer was later used to generate data for model training.

Entering Cloud Storage Era: the ultimate elasticity of data service, 2022 (slides)
Tail, Walk with Me (a comprehensive guide to tail latency), 2022 (slides)
Adapting TiKV for Cloud Storage, 2022 (CNCF)
Raft Engine in TiKV 6.1, 2022 (slides, InfoQ)
Best Practices for TiDB on AWS Cloud, 2020 (blog, DZone)
Sightseeing OceanBase’s Storage Design, 2020 (slides)

Zhejiang University - Computer Science (Bachelor; 2016 ~ 2020)