Want to know if this job is worth applying to?
San Francisco, California, United States
Onsite
Arena is the platform for evaluating how AI models perform in the real world. Founded by researchers from UC Berkeley's SkyLab, we're on a mission to measure and advance the frontier of AI for real-world use, and to build the foundation for everyone to understand, shape, and benefit from it.
Tens of millions of people use Arena each month to evaluate how frontier systems handle the work they actually do. The preferences they share power the most transparent, rigorous, and human-centered evaluations in AI. Leading AI labs, enterprises, and independent researchers rely on our work and open datasets to understand how models behave in real workflows: agentic coding, creative generation, professional productivity, and beyond. We go beyond leaderboards and decompose what human experience reveals about AI, so models advance toward the work people actually do.
We're a team of researchers, academics, builders, and creatives from UC Berkeley, Google, Stanford, and DeepMind. We seek truth, move fast, and value craftsmanship, curiosity, and impact over hierarchy. We're building a company where thoughtful, curious people from all backgrounds can do their best work together, in an office culture that radiates excellence, energy, and focus.
Arena Intelligence is seeking a Software Engineer - Anti-Abuse to own platform misuse end-to-end. Arena's evaluations are only as trustworthy as the signal behind them — and that signal is under constant, creative attack. You will build the detection, enforcement, and investigation systems that keep Arena's leaderboards trustworthy, stop automated abuse across our services, and defend against the full spectrum of AI-era harms.
This is a founding builder role. You will set the strategy, write the code, and build the platform that future abuse, integrity, and trust & safety hires grow on top of. You'll work shoulder-to-shoulder with product, infrastructure, model partners, policy, and leadership, and you'll be accountable for outcomes the whole company can see: is the leaderboard clean, are harmful uses caught, are our services safe to ship?
Own the abuse vision for Arena: what gets detected, what gets enforced, how fast, and with what false-positive budget
Design and operate detection for bots, sybils, coordinated inauthentic voting, and rating-system manipulation — the integrity of Arena's leaderboards is the product
Build enforcement primitives (rate limits, challenges, shadowbans, account actions, model-side refusals) that are reversible, auditable, and humane
Detect and mitigate inference abuse and cost exploitation at the platform layer
Build jailbreak and multi-provider misuse detection across the models Arena serves, and partner with model-provider trust & safety teams on signal-sharing and escalation
Scope and implement abuse monitoring for every new product launch — web search, web fetch, live site deployment, and whatever's next — as part of the launch checklist, not after the fact
Prototype and mature into production systems of detection, review, and enforcement for the highest-severity harms (CSAM/NCII, violent extremism, self-harm), including the legal reporting pipeline (e.g., NCMEC)
Build internal investigator tooling so policy, on-call, and future T&S analysts can triage incidents without engineering bottleneck
Partner with Security on shared surface — account takeover, credential stuffing, API-key abuse, and the identity/behavioral-signal platform
Partner with policy, legal, and leadership on acceptable-use policy, enforcement escalations, and public-integrity narrative
6+ years of production software engineering experience, including building and operating systems under adversarial conditions
Shipped experience in at least one of: trust & safety, anti-abuse, anti-fraud, anti-spam, integrity, or risk engineering
Strong SQL and data-analysis skills — this role is 30%+ pattern-finding and investigation, not just shipping code
Adversarial and investigative mindset — you can articulate a novel attack before designing the defense, and follow evidence when a novel harm surfaces
High judgment on false-positive cost, user harm, and the reversibility of enforcement actions
Proficiency in a modern backend language (Node.js, TypeScript, Python, or Go)
Excellent communication — you'll build alignment with engineering, product, policy, and leadership routinely
Bonus Points
Experience with LLM-specific adversarial inputs — jailbreaks, direct and indirect prompt injection, tool-use abuse
Experience with agent safety, browser-automation abuse, or LLM-API abuse
Background in securing voting, rating, reputation, or marketplace platforms against coordinated manipulation
ML or ML-systems experience — feature engineering, online/offline evaluation, label acquisition, drift handling
Experience building investigator or analyst tooling used by non-engineers
Contributions to open-source trust & safety, abuse-detection, or adversarial-ML work
Background in gaming integrity, ad-fraud, or financial-crime engineering at scale
We offer competitive compensation and equity aligned to the markets where our team members are based. The base salary range will depend on the candidate’s permanent work location.
Comprehensive health and wellness benefits, including medical, dental, vision, and additional support programs.
The opportunity to work on cutting-edge AI with a small, mission-driven team
A culture that values transparency, trust, and community impact
Come help build the space where anyone can explore and help shape the future of AI.
Arena Intelligence provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, national origin, age, disability, genetics, sexual orientation, gender identity, or gender expression. We are committed to a diverse and inclusive workforce and welcome people from all backgrounds, experiences, perspectives, and abilities.
