5 yrs in AdTech. At InMobi led the delivery engine (retrieval → ranking → bidding → delivery); owned frequency capping, budgeting, callback/attribution integration, core components and performance tuning. At Zhaolan built DSP/DMP. Deep expertise across the full RTB bidding and oCPX bid pipeline.
Performance & Algorithms
Ad engine handling 100K+ QPS; optimized P99 to 60ms, cut compute 50–60%+ and storage 40%+. Deployed FM/FFM via TensorFlow at <2ms per inference; built A/B experimentation and monitoring
Agentic Engineering
Since early 2026, built Agentic Engineering workflows with Claude Code + multi-model collaboration covering coding → testing → review → PR, with auto Jira task creation and task-driven code analysis (each step human-confirmed). From Vibe Coding to Agentic Engineering — AI accelerates delivery while architecture, business judgment and quality stay human-led. Master the tools, don’t be mastered by them
Cross-domain Engineering
AdTech engines, blockchain core (uni-chain / contracts / DEX), and financial compliance (KYC/AML). Built Go microservice stacks with Docker + Consul + Prometheus/Grafana
Tech Stack
Languages
Go · Java · Python · JavaScript · Bash
Architecture
Gin · gRPC · Protobuf · Spring Boot · React
Databases
MySQL · PostgreSQL · Redis · Elasticsearch · ClickHouse
Contributed to the compliance engine (risk scoring, policy configuration, automated review); built Agentic Engineering workflows to boost cross-stack delivery
Independent Developer
Full-stack Development
2024.02 – 2025.11
Freelance technical project development and delivery.
Took on Go/Java backend and full-stack projects; helped clients solve various engineering problems
InMobi
Senior Golang Engineer · Demand Team · Beijing
2020.01 – 2024.01
A leading global mobile AdTech platform operating across APAC, North America, Europe and beyond.
Ad Delivery Engine: led core modules (request handling, retrieval, ranking, bidding, delivery, frequency capping, budgeting, callbacks); production load 100K+ QPS
Performance: cut P99 from 80–100ms to 10–60ms (connection/memory pools, lock optimization, GC tuning); reduced compute 50–60%+ and storage 40%+
Model Inference:Deployed FM/FFM (CTR/CVR prediction) via TensorFlow for online inference + batch estimation at <2ms each; built the A/B experimentation framework
Infrastructure: Go microservice stack (Consul + Docker + ELK + Prometheus/Grafana); ran code reviews and tech-sharing sessions
Multi-strategy valuation service supporting market-price, residual-value and historical-transaction pricing models; built a simulation/replay system (with web UI) from scratch for model backtesting
Go + Docker two-stage builds, Consul service discovery, ELK logging, Kafka + Druid data pipeline
Zhongsiboan Technology
Software Engineer · Beijing
2016.08 – 2018.09
Focused on an in-house blockchain core platform, smart-contract engine and decentralized apps.