Principal Machine Learning Engineer - Matchmaking
Roblox · San Mateo, CA, United States
About this role
**Principal Machine Learning Engineer — Matchmaking (Roblox)** Roblox is building the tools and platform that empower our community to bring any experience to life—connecting a billion people with optimism and civility. **About the role** Matchmaking is a critical stage downstream of Discovery and Search. After a player chooses a game, matchmaking assigns them to a server—shaping who they play with, the network latency they experience, and the quality of their session. As a **Principal ML Engineer on Matchmaking**, you’ll own the technical vision and execution for the models, objective functions, and algorithms that drive these decisions at massive scale. You’ll combine predictions of high-value player actions with optimization of player/developer utility and system efficiency, serving millions of players across hundreds of thousands of servers in global datacenters. **You will:** - Own the technical direction and ML strategy for matchmaking, connecting Discovery/Search to server placements that improve player experience, developer success, and Roblox growth. - Define and evolve the matchmaking objective function to balance player and developer utility with system efficiency—making tradeoffs explicit across co-player compatibility, network/join latency, engagement/retention, server utilization, and compute cost. - Build models that predict high-value actions and longer-term outcomes for each player–server match (e.g., co-play, chat engagement, sustained play, return visits) using player history, game characteristics, and live server state. - Design and productionize candidate-server retrieval, ranking, and assignment algorithms that optimize the objective under real-time latency, capacity, and reliability constraints. - Lead by example as an individual contributor: mentor engineers and build an engineering culture of excellence. - Build and evolve data, feature, training, serving, and evaluation systems to support matchmaking ML with fresh signals and consistency between training and serving. - Partner with Discovery, Search, developer-facing product teams, modelers, and data/AI infrastructure engineers to co-design systems that adapt to each game’s needs. - Define offline evaluation and lead online experiments to measure impact on engagement/retention, developer success, and system efficiency—then iterate models and objectives. - Translate research into production improvements and represent Roblox’s technical work. **You have:** - 8+ years of experience delivering and improving large-scale ML systems for ranking, recommendation, personalization, or resource allocation. - Experience defining objective functions, building predictive models, and deploying optimization algorithms that balance user/business outcomes with production latency, reliability, and resource constraints. - Ability to break down long-term product requirements into iterative deliverable stages. - Experience initiating and facilitating challenging technical decisions across teams; empathetic and able to build shared understanding. - Deep technical expertise with a user-first mindset (Roblox users and internal engineers). - Generalist strengths—comfortable across multiple languages/technologies and adapting quickly. - Willingness to roll up your sleeves to tackle data and ML problems as they arise. - Creative, strategic problem-solving; ability to simplify the complex and bring fresh ideas. - BS in Computer Science (or related). MS/PhD is a plus. **Location / sche
Listing freshness
CronJobs last confirmed this listing 1h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.