# Huayang Guo > Build the next generation of intelligent products that combine data, AI, and growth. Location: San Francisco, California, United States Profile: https://flows.cv/huayang Product engineer with deep expertise in AI-driven product growth and large-scale system design. Experienced in building end-to-end pipelines from web/mobile frontend to backend, data infrastructure, and experimentation frameworks. Proven track record in applying LLMs, generative AI, and intelligent recommendation systems to boost user acquisition, engagement, and feature adoption. ## Work Experience ### Staff Software Engineer @ Snowflake Jan 2024 – Present | Menlo Park, California, United States • New User Acquisition – Own the full free-trial signup and conversion pipeline (UI, backend services, sales/self-serve integration). • AI-enhanced Feature Adoption – Lead the design and implementation of Winterfest, an internal promotion platform delivering personalized feature recommendations powered by predictive models. Integrated targeting algorithms that adapt in real time to user behavior, increasing promotion-driven feature usage by 50–100%. • Established AI-assisted engineering workflows, standardizing LLM-based coding, automated test generation, and documentation creation. • Built an internal AI-driven dashboard tool leveraging RAG on high-quality past queries to accelerate data analysis and query generation for North Star metric tracking. • Partner with PMs, data scientists, and engineers to define North Star metrics, run controlled experiments, and scale successful strategies. ### Senior Software Engineer @ TikTok Jan 2024 – Jan 2024 | San Jose, California, United States • Built a chat-based creation interface in Effect House, allowing creators to design custom video effects with natural language prompts powered by GPT-based language models. • Designed scalable backend APIs for effect rendering and processing in TikTok’s effect backend. • Contributed to model evaluation and prompt engineering strategies to balance creativity, safety, and performance in production deployment. ### Software Engineer @ Coda Jan 2020 – Jan 2024 | San Francisco Bay Area Believe that a "doc" should be redefined. Key achievements: - Designed and implemented secure, isolated execution environments for third-party code. - Scaled backend infrastructure to support public usage and platform growth. - Led the authentication system migration and significantly improved system performance. - Authored the developer protocol and SDK used by external builders to create Packs. - Contributed across the stack, from frontend UI integration to backend APIs and infrastructure. ### Staff Software Engineer @ Pinterest Jan 2018 – Jan 2020 | San Francisco Bay Area Build ads products. ### Senior Software Engineer @ Pony.ai Jan 2018 – Jan 2018 | Fremont, CA ### Cofounder and CTO @ Shoppo Jan 2017 – Jan 2018 | Shanghai City, China Built an eCommerce platform that sells made-in-Asia products to US consumers. - infrastructure on AWS, plus optimization for sellers from China - mobile stack purely on React Native - web on React - major API layer on GraphQL/python app store https://itunes.apple.com/us/app/shoppo-shop-the-world/id1276073764?mt=8 ### CTO @ Operator Jan 2017 – Jan 2017 | Shanghai / San Francisco Leading the eng org and China expansion. Operator is focusing on the cross-border e-commerce that makes authentic US goods available for shoppers all over the world. Please join us to build the strong relationship with both US partners and China consumers. Both SF and Shanghai positions are available! ### Software Engineer @ Operator Jan 2016 – Jan 2017 | San Francisco Brought up the China project which opens up a brand new & super promising market for the company. - localized the US app into China app (including i18n and localized features like payments/shipping) - set up the cross-border infrastructure to ensure app runs smoothly while the servers are ocean distance away. - built up the China eng team. - worked on every piece of the tech stack. ### Software Engineer @ Pinterest Jan 2014 – Jan 2016 | San Francisco Bay Area Project: Monetization products. Pinterest started monetization in 2014. With a 4-people team, we were able to build first version of ads.pinterest.com in 6 weeks, V2 of analytics.pinterest.com in 2 months, ads API in another 3 months. Today, these platforms account for all ads creation on Pinterest. I specifically worked on - Major code contributor to web, API, datalayer service, dashboards, analytics system (HBase) optimization Project: Pins. Pin is the most important concept for Pinterest. We are making it more useful by adding more information to a Pin, collecting and sharing user-added data on Pins, etc. I specifically worked on: - Re-implementing Place Pin at Pinterest, increasing its traffic by 200x to cover ~10% of Pins - Building Canonical Pin product which groups Pins by image and link, and shares useful information between Pins in the same group. It’s expected to double board followers. - Initiated the make-card product on Pin, which introduces well-formatted DIY content to Pinterest. Building a reader view experience (like Readability) in the in-app browser to improve the click through experience from Pins. ### Software Engineer @ Facebook Jan 2012 – Jan 2014 | Menlo Park 1. Cross-datacenter cache consistency: about: Facebook has multiple datacenters, each of which runs a caching cluster on its own. In order to serve user requests to all datacenters, a hot piece of data can be replicated in all the caching clusters. Upon mutating/deleting, it may happen that the piece of data varies in different datacenters. The project aims at detecting this kind of cache inconsistency and helping debugging the root cause. my role: participated in building the monitoring system and tools, tracked down engineering issues that made cache inconsistent. 2. Cross-datacenter cache invalidation end-to-end testing: about: cache invalidations are conveyed on a per-shard basis. the hundreds of thousands of shards in FB make it hard to detect a small number of stuck shards (network partition, software error, etc). however, it hurts since users on the stuck shard will see badly stale cache. the testing system aims monitoring invalidation on a regular and per-shard basis. it has helped detecting and preventing several major stale cache issues in 2013. my role: designed and built the automatic testing system, built alerts on top of it and built UI debugging tools that helps engineers diagnose root cause of invalidation failure. 3. ads delivery estimation: about: the project aims at answering the following question(s). - when you create an ad on Facebook, how you'd expect your ad to deliver in terms of impressions/revenue/etc. - when you update a running ad, how you'd expect the updated ad to delivery differently in terms of impressions/revenue/etc. this is hard because the new/updated ad hasn't be delivered yet. we try to estimate this problem by simulating the problem on a sampled set of users with the major component of ads delivery system (e.g. auction). my role: re-designed and re-built the accuracy evaluation pipelines and dashboards, implemented several optimizations or bug fixes that improved the accuracy confident level from 30% to ~80%. ### Research Intern @ Microsoft Research Asia Jan 2009 – Jan 2011 Model checking and related optimization. ## Education ### Master of Science (M.S.) in Computer Science Columbia University Jan 2011 – Jan 2012 ### Bachelor of Engineering (BE) in Computer Science Tsinghua University Jan 2007 – Jan 2011 ## Contact & Social - LinkedIn: https://linkedin.com/in/huayang-guo-profile - Website: http://huayang.me --- Source: https://flows.cv/huayang JSON Resume: https://flows.cv/huayang/resume.json Last updated: 2026-03-22