About me
10+ years in big data with 5+ years in risk control. I have spent years researching and building in big data, distributed systems and risk control, and I enjoy turning hard technical problems into shippable engineering.
I'm an Apache HugeGraph Committer and keep refining our internal tech stack through open source. This site is both my online resume and a place to distill my thinking — feel free to reach me via the channels below.
Experience
2020 — Present
Lalamove (Huolala)
Big Data Expert · Business Risk ControlBeijing- Owned big-data & graph tech for the risk-control dept.; self-built real-time graph storage cut subsidy loss rate from 10% to under 1%.
- Delivered real-time graph services for freight, mobility & international; graph strategies/features for brush, fake-order, content-safety & order-switching risks.
- Evolved graph storage: HBase middleware (k-hop P99 ≤ 100ms) → adopted HugeGraph for Gremlin real-time queries, shortening delivery from weeks to hours.
- Built user/driver entity graphs from device fingerprints & payment IDs; Spark offline mining of fake-newbie, large-subgraph & same-person labels powering subsidy & watchlist 1-hop.
- Unified graph modeling language (Gremlin) & requirement workflow, cutting delivery cycle and collaboration cost.
- Open source & impact: Apache HugeGraph Committer (kernel optimization, pre-partitioning); hackathon 4th place; LLM-based Graph Agent for self-serve analysis.
2018 — 2020
Zhihu
Data Platform EngineerBeijing- Built OLAP engine on Druid for company-wide real-time BI; 100+ nodes, 1PB+ storage serving growth & monetization dashboards.
- Stability: server-side query timeout & resource release kept 99% of queries under 10s.
- Evaluated ClickHouse / Doris and delivered POCs; authored Zhihu column "Druid Cluster Optimization Practice".
2017 — 2018
Didi Chuxing
Senior R&D Engineer · Infra PlatformBeijing- Rebuilt HBase monitoring (collect / store / visualize), cutting MTTR and improving early detection for key businesses.
- User support: RowKey design & capacity planning; resolved data skew/hotspot, RS CPU spikes and crashes.
2015 — 2017
Saisi XinAn
Big Data EngineerBeijing- Designed & built an HBase-based small-file system for S3-like storage.
- Deployed 100-node cluster ingesting 200M files/day with 140TB / 7-day retention.
Tech stack
Distributed & Big Data
Graph & Risk Control
Warehouse & OLAP
Engineering & Languages
Education
Harbin Institute of TechnologyB.S. · Software Engineering
Class of 2015Honors & Awards
2023Apache HugeGraph Committer
2024Lalamove Hackathon — 4th place (team lead)
×4Lalamove Tech-Center Value Star
3 yrsTop 15% performance (S / A), Risk Control