Principal ML Engineer - Matchmaking & Ranking NEW

The Briefed; San Mateo (CA); On-site; USD 295,000 - 345,000; Full time

Benefits offered by this job

Healthcare & Wellness

401(k) Retirement Savings

HQ Campus Perks

Job summary

Roblox is seeking a Principal ML Engineer on Matchmaking to own the technical vision for ML models, objective functions, and ranking systems that govern server placements for players’ experiences. You will lead predictive modeling, real-time retrieval, and offline/online evaluation to balance latency, player utility, and compute costs.

The role emphasizes cross-functional leadership, mentoring engineers, and translating cutting-edge ML research into production gains across Discovery, Search, and

Qualifications

• 8+ years of professional experience delivering and scaling production ML systems for ranking, recommendation, personalization, or resource allocation.
• Deep expertise defining multi-objective loss functions, building predictive models, and deploying optimization algorithms under latency/resource constraints.
• Hands-on experience building end-to-end MLOps workflows (feature stores, real-time inference, training pipelines, evaluation).
• Mastery across multiple languages (Python, C++, Go, Java, or Luau) for data pipelines and infrastructure code.
• Ability to steer long-term product goals into iterative engineering milestones across teams.
• Bachelor’s degree in Computer Science, Software Engineering, ML, or related field (Master’s/PhD preferred).

Responsibilities

• Own the technical vision and ML strategy for matchmaking, connecting Discovery and Search to intelligent server placements.
• Formulate multi-objective utility functions balancing player compatibility, latency, join times, session length, retention, and compute costs.
• Build and deploy models predicting high-value player actions and long-term engagement using historical data and live state.
• Design candidate retrieval, ranking, and server-assignment algorithms with real-time latency constraints.
• Evolve ML infra, feature stores, training/serving pipelines, and online inference to minimize training-serving skew.
• Lead offline benchmarking and online A/B experiments to measure impact on retention and platform efficiency.
• Mentor senior engineers and align architecture with Search, Discovery, Data/AI Infra, and Product teams.

Skills

ML engineering

Optimization

Distributed systems

Programming languages

Cross-functional leadership

Education background

Education

Bachelor's degree in CS/SE/ML

Job description

Every day, tens of millions of people come to Roblox to explore, create, play, learn, and connect across 3D immersive digital experiences. Matchmaking is a critical stage downstream of Search and Discovery—once a player selects an experience, matchmaking assigns them to a server in a global datacenter. That decision dictates co-player compatibility, network and join latency, session quality, and long-term player retention.

As Principal ML Engineer on Matchmaking, you will own the technical vision, strategy, and architecture for the machine learning models, objective functions, and optimization algorithms that govern these decisions. Operating as an individual contributor lead, you will build predictive models for high-value player actions (co-play, sustained play, return visits), design real-time candidate retrieval and ranking systems, and balance complex trade-offs between player utility, developer success, server utilization, and compute cost across millions of concurrent players.

Key Responsibilities

• Technical Strategy & ML Direction: Own the technical vision and machine learning strategy for matchmaking, connecting Discovery and Search user journeys to intelligent server placements.
• Objective Function Optimization: Formulate and evolve multi-objective utility functions that explicitly balance player compatibility, network latency, join times, session length, retention, server utilization, and compute costs.
• Predictive Model Development: Build and deploy models that predict high-value player actions and long-term engagement outcomes using historical player data, game metadata, and live server states.
• Real-Time Retrieval & Ranking Infrastructure: Design candidate retrieval, ranking, and server-assignment algorithms that execute under tight real-time latency, capacity, and reliability constraints.
• Data & Feature Engineering: Evolve ML infrastructure, feature stores, training/serving pipelines, and online inference systems to eliminate training-serving skew and deliver low-latency predictions.
• Offline & Online Evaluation: Lead offline benchmarking and online A/B experimentation to measure the impact of matchmaking algorithms on player retention, developer monetization, and platform efficiency.
• Cross-Functional Leadership: Mentor senior engineers, drive architectural alignment with Search, Discovery, Data/AI Infra, and Product teams, and translate cutting-edge ML research into production gains.

Requirements

Machine Learning & Distributed Systems Engineering Depth

• 8+ years of professional experience with a track record of delivering and scaling production ML systems for ranking, recommendation, personalization, or resource allocation.
• Objective Functions & Optimization: Deep expertise defining multi-objective loss functions, building predictive models, and deploying optimization algorithms under strict latency and resource constraints.
• High-Scale Production Engineering: Hands‑on experience building end‑to‑end MLOps workflows—including feature stores, real‑time inference, model training pipelines, and evaluation loops.
• Polyglot Systems Capability: Mastery across multiple programming languages (Python, C++, Go, Java, or Luau) with the ability to navigate data pipelines and low-level infrastructure codebases.
• Cross-Functional Influence: Proven ability to break down ambiguous long‑term product goals into iterative engineering milestones while facilitating alignment across multiple engineering teams.
• Education: Bachelor’s degree in Computer Science, Software Engineering, Machine Learning, or a related quantitative field (Master’s or PhD preferred).

Preferred Qualifications

• Prior experience with real‑time matchmaking algorithms, game server allocation, or dynamic multi‑tenant cloud scheduling.
• Deep expertise in reinforcement learning (RL), contextual bandits, or graph neural networks (GNNs) applied to social graphs and co‑play compatibility.
• Active contributions to the broader machine learning research community (e.g., NeurIPS, KDD, RecSys papers or open‑source libraries).

Total Rewards & Benefits

• Compensation & Equity: Annual base salary range of $295,250 – $345,040 USD + comprehensive equity compensation package.
• Healthcare & Wellness: Comprehensive medical, dental, and vision insurance options for employees and eligible dependents.
• Retirement & Financial: 401(k) retirement savings plan with employer match, life insurance, and flexible spending accounts.
• HQ Campus Perks: On‑site catered meals, modern collaborative workspaces, learning and development stipends, and team events.