Senior Software Engineer (User Behavior)
Posted Updated
Description
- You will build the backend systems that fundamentally shift how Roblox responds to unsafe behavior - from reactive, stateless punishment to a personalized, adaptive system that reflects who a user actually is
- You will own core backend infrastructure for a new capability at Roblox: a low-latency behavioral data layer, a centralized rules-and-inference engine, and a unified intervention-delivery framework that together let us treat users based on their real risk profile instead of a single point total
- This is a rare opportunity to build foundational, high-scale systems that directly shape how Roblox understands and responds to risk across hundreds of millions of users
- You will feel a deep sense of responsibility building systems that directly affect millions of users, especially minors, on one of the world’s largest platforms
- Your work will reduce harm, improve fairness, and help ensure Roblox remains a place where creativity and community can thrive safely
- Build the User Behavior Foundation: Design and operate the low-latency, per-user behavioral data layer that powers real-time safety decisions - aggregating signals like violation history, in-game reports, and behavioral patterns into a single queryable source of truth
- Own the Consequence Engine: Build and evolve the centralized decision engine that combines deterministic rules with ML risk signals into a single, consistent safety decision the authoritative system every consequence flows through
- Own the Feature Restriction Platform: Build a unified, backend-driven framework for delivering safety interventions that any product team can plug into, replacing today’s fragmented, feature-specific implementations
- Unify Abuse Reporting: Build the backend that consolidates reporting channels into a single, scalable framework - integrating LLM-based classification, agentic review workflows, and ML models to improve report routing, prioritization, and reporter feedback while meeting global regulatory requirements
- Design for trust and auditability: Build serving infrastructure that meets a high bar for explainability and fairness, since these systems directly determine consequences for millions of users
- Ship code, not just designs: Own end-to-end delivery - from system design through migration, testing, and production operation
- 5+ years of experience building and operating production backend systems at scale, ideally in Go or a comparable statically-typed backend language
- Solid production operations fundamentals: monitoring, alerting, oncall, capacity planning, and incident response for systems with real user-facing impact
- Comfortable with Ambiguity: this space is actively being redefined — you’ll help set the standard, not just implement one handed to you
- Committed to Ethical, Responsible Engineering: dedicated to building systems that are fair, auditable, and operate with the utmost responsibility toward user safety
- Nice to have: experience in Trust & Safety, fraud, risk, or other adversarial domains; experience with feature stores or ML-serving infrastructure; experience building frontend-agnostic, backend-driven UI frameworks (e.g. server-driven intervention/notification rendering)
- Experience with streaming or batch data pipelines feeding a feature store or similar ingestion layer, and comfort partnering with ML engineers who consume that data
- A Reliability-Minded Engineer: you treat testing, migrations, and oncall as first-class parts of the job, not afterthoughts
- A Pragmatic Builder: scrappy and impact-oriented, comfortable turning scattered, undocumented signals and legacy code paths into clean, well-owned services
- A Systems Thinker: capable of designing infrastructure that other teams (ML, feature teams, policy) build on top of, not just point solutions
- Experience building and evolving rules engines, workflow/decision systems, or similarly stateful business-logic services
- Strong experience designing distributed systems: service APIs (RPC/gRPC), data modeling for high-throughput event ingestion, caching, and low-latency read paths