Shu Zhang

CTO

Palo Alto, California, United States12 yrs 3 mos experience

Highly StableAI Enabled

Key Highlights

Scaled Pinterest's user base and revenue significantly.
Led the development of advanced AI systems.
Founded and managed high-performance engineering teams.

Stackforce AI infers this person is a SaaS expert with a strong focus on AI and machine learning infrastructure.

Contact

Skills

Core Skills

Machine LearningAi SystemsEngineering ManagementInfrastructure Engineering

Other Skills

PersonalizationAgent ApplicationsRecommendationsQuery UnderstandingBig DataGPU ServingMLOpsFeature EngineeringTeam LeadershipAI InnovationSystem ArchitectureC++Distributed SystemsPythonJava

About

Dedicating my career to building high performance teams and AI systems with my experience as an early member at Pinterest and Perplexity, where I helped scale Pinterest with 10x valuation, 10x user base and 500x revenue since 2014. At part time I did some angel investments to help early stage companies grow.

Experience

12 yrs 3 mos

Total Experience

11 yrs

Average Tenure

1 yr 3 mos

Current Experience

Perplexity

Member of Technical Staff, Head of Applied AI

Jan 2025 – Present · 1 yr 3 mos

We are hiring at San Francisco, Palo Alto, and New York!
DM me if you are interested in following areas, or any exchange of ideas is always welcomed:
Personalization
Agent Harness and Platform
SFT and RL post training
Agent Applications
Recommendations
Eval and Data Flywheel
Query Understanding

Machine LearningPersonalizationAgent ApplicationsRecommendationsQuery UnderstandingAI Systems

3 roles

Director of Engineering

Promoted

Jan 2023 – Jan 2025 · 2 yrs · San Francisco Bay Area

Support the Ads ML Infra team to accelerate the multi-generational AI/ML application transformations. Scaled the team from 4 to 40.
Ads ML Serving infra: Highly scalable ads ML inference systems and ads prediction framework. GPU serving for large models and embedding tables.
Ads ML Training Infra: Big data systems, GPU training for transformer models, training data lake, user-friendly MLOps toolkits, last mile data processing.
Ads ML Feature Infra: Unified feature engineering platform (logging, serving and generation). Large text and sequence features. Signal loss and privacy preserve ML features.

Machine LearningBig DataGPU ServingMLOpsFeature EngineeringEngineering Management

Engineering Manager

Jan 2019 – Jan 2023 · 4 yrs · San Francisco Bay Area

Founder of Ads ML Infra team. We are a group of talented engineers with a mission of accelerating AI innovation velocity and quality for Pinterest Ads by providing reliable, efficient, user-friendly and unified ML systems.

Machine LearningTeam LeadershipAI InnovationEngineering Management

Staff Software Engineer

Jan 2014 – Jan 2019 · 5 yrs · San Francisco Bay Area

2017 - 2019 | Tech Lead, Ads Serving Platform. Built low-latency, scalable, reliable and debuggable infrastructure to deliver high quality ads to pinners. Led a major re-architecture of ads serving infra and eliminated several bottlenecks for serving 100x ads inventory.
2015 - 2017 | Senior Engineer, ML Serving Platform. Built high performance ML systems to serve hundreds of models and score tens of millions of objects per second. Author of Pinterest's first model management library and first ML online inference platform.
2014 - 2015 | Service Framework. Built multi-language framework and toolkits to facilitate SOA within Pinterest. Also owning image processing pipelines. Author of Pinterest's first realtime configurator system.

Machine LearningInfrastructure EngineeringSystem Architecture