Boyu Fang
Senior Software Engineer · Recommender Systems & ML Infrastructure
Experience
Meta New York, NY · Jun 2022 – Present
Senior Software Engineer
Instagram Relevance Labs — Founding engineer & tech lead, “Tune Your Algo” (2024–Present)
- Designed and built the first end-to-end prototype of Tune Your Algo and defined the long-term architecture for user-controllable recommendations on Instagram.
- Led a cross-functional team of ~10 engineers across ranking, infra, and product from incubation to large-scale rollout.
- Shipped Tune Your Algo to all English-speaking users (~1.7 billion people) — Meta’s first AI-powered system letting users explicitly steer the recommendation algorithm at scale.
- Drove alignment with senior leadership, securing investment and headcount via the product vision, technical strategy, and projected impact.
- Established internal best practices for user-controllable recommendation systems (controls, telemetry, safety), used as a reference by other relevance teams.
Instagram — Reposts ranking & infrastructure (H1 2024)
- Led core ranking and infrastructure for Instagram reposts as part of new “Friends” experiences, contributing to nearly 1% core session growth on IG, especially among teens.
Instagram — Feed Relevance infrastructure (2022–2023)
- Designed and implemented an internal ranking DSL and interpreter used across Instagram Feed Relevance.
- Optimized feature fetching during inference and training-data logging, reducing latency and infra cost.
- Cumulatively saved >1000 ms of latency and >1 MW of power across Instagram Feed infrastructure.
- Improved sourcing and delivery components of the feed system for better efficiency and reliability.
Featured in The Verge, Mashable, ABC News, and Meta Newsroom; highlighted by Adam Mosseri (Head of Instagram).
TikTok London, UK · Mar 2021 – Jun 2022
Senior Software Engineer, Business Integrity
- Redesigned the policy search tool to surface the most relevant policies per query — 90% of user clicks landed on the first page of results.
- Built a data-driven policy iteration process with policy insights and policy-simulation A/B testing tools.
- Led a major refactor of the policy-management backend for reliable, low-risk policy updates.
Quant Dev & Market Infra London, UK · Aug 2015 – Feb 2021
Quant / Senior Developer — Bloomberg, Citadel, Millennium, Squarepoint
- Squarepoint — Built a U.S. power-grid simulation platform letting quants run tens of thousands of FTR scenarios per day to support trading decisions.
- Millennium — Built from scratch a macro trading research and production stack (signals, portfolio construction, backtesting) for a new systematic team, supporting hundreds of strategies in production.
- Citadel (Global Fixed Income) — Implemented a real-time live P&L framework, cutting production incidents ~50% and improving P&L latency for traders and risk.
- Bloomberg (FXGO) — Engineered core components of Bloomberg’s FXGO institutional FX trading platform, focused on low latency and reliability.
Education
New York University, Center for Data Science
Part-time M.S. Data Science (in progress)
2025 – Present Part-time M.S. Data Science (in progress)
University of Cambridge
MPhil Advanced Computer Science
2014 – 2015 MPhil Advanced Computer Science
University of Bristol
BSc Computer Science, First Class Honours
2011 – 2014 BSc Computer Science, First Class Honours
Open Source & Side Projects
- duibiao.info — Co-maintainer of a transparency site for the Chinese tech job market.
- pytomlpp — Maintainer of a Python library wrapping a C++ TOML parser for fast TOML processing.
Skills
- Languages
- C++, Python, TypeScript, JavaScript
- Domains
- Large-scale recommender systems, ranking infrastructure, ML pipelines, experimentation, LLM applications, quant trading
- Cloud & Infra
- AWS, Google Cloud, Docker, distributed systems
- Tooling
- GitHub, GitLab, CI/CD
- Spoken
- English, Chinese