Jobs at Alice (Formerly ActiveFence)

Principal Data Scientist leading the technical direction of GenAI safety, content-moderation inference, red-teaming, evaluation, and MLOps. The role owns production-scale ML architecture across Python, Rust, PySpark, Databricks, Triton, and MLflow; leads ambiguous projects from framing through rollout; evaluates LLM quality, safety, latency, and cost; and mentors data scientists and ML engineers while partnering with Product and customers.

Own the AgenticVerse production SaaS platform across frontend, backend, data models, APIs, workflows, reliability, access control, and observability. Build realistic digital twins with user interfaces, APIs, persistent state, and verifiable event streams. Collaborate with researchers to deliver features from design through production. The role requires strong TypeScript, React, backend development, API and data modeling, Docker, cloud infrastructure, and high autonomy.

Build and own adversarial reinforcement-learning environments, including replica systems, planted attacks, verifiers, task evaluation, and scalable run infrastructure. Design secure, sandboxed environments for tool-use, web, and computer-use agents; analyze rollout traces to prevent reward hacking; and package environments for customer training pipelines. Collaborate with researchers, review designs and code, and improve task quality for frontier AI model training.

Lead customer-facing technical programs for frontier AI labs, including reinforcement learning environments, training tasks, evaluation infrastructure, automated rewards, and verification systems. Manage multiple programs, high-volume task production, external technical contributors, cybersecurity and agent safety initiatives, and cross-functional delivery across research, engineering, product, sales, and customers.

Lead multiple technical programs developing reinforcement learning environments, training tasks, evaluation infrastructure, automated rewards, and verification systems for frontier AI models. Translate research and customer requirements into roadmaps and scalable delivery workflows. Manage distributed technical contributors, cybersecurity and agent-safety evaluations, and high-volume task production while coordinating Research, Engineering, Product, Sales, customers, and external experts.

Research emerging security risks in AI agents, including prompt injection, tool misuse, data exfiltration, memory poisoning, privilege escalation, and unsafe autonomy. Build reproducible experiments, sandboxes, detectors, graders, and regression tests across enterprise AI stacks. Translate findings into platform guardrails, red-team tests, threat intelligence, customer evidence, technical publications, conference talks, and responsible vulnerability disclosures. Collaborate with product, engineering, marketing, and external AI research organizations.

Lead Alice Labs’ AI safety and security research agenda, covering model evaluation, red teaming, adversarial robustness, LLMs, agents, and multimodal systems. Own the research roadmap, design industry-leading benchmarks, publish at top-tier conferences, mentor senior researchers, manage academic and industry partnerships, and translate research into products. The role also represents Alice externally at conferences, partner organizations, and frontier labs, requiring significant international travel.

Leads product strategy and execution for Trust & Safety and security transformation, converting bespoke moderation workflows into scalable, data-driven products. Owns agentic moderation roadmaps, data assets, Big Tech platform requirements, productization, and bespoke project delivery. Partners with cross-functional teams to develop autonomous investigation and enforcement systems that operate at enterprise scale.

Lead product strategy and execution for Trust & Safety and security transformation, turning bespoke moderation workflows into scalable, data-driven products. Own the roadmap for agentic moderation systems using LLM-driven agents for autonomous investigation and enforcement. Partner with domain leaders to productize expertise, maintain a durable data moat, scale offerings for major technology platforms, and deliver selected customer projects from requirements through production.

Manage global HRIS operations, employee lifecycle updates, onboarding and offboarding, data integrity, compliance audits, recurring HR reporting, compensation and benefits data, and AI-enabled HR automations. Support HRBPs and Talent Acquisition teams while collaborating with Legal, Finance, IT, and business stakeholders. The role also maintains HR platforms, benchmarks, employee records, and analytics workflows.

Leads domain experts across penetration testing, red teaming, cloud security, and ICS/OT security. Responsibilities include assigning work, balancing capacity, monitoring quality and throughput, resolving cross-domain edge cases, maintaining annotation guidelines, and reviewing prompts directly. The role requires deep hands-on expertise in at least one cybersecurity domain, working knowledge across all covered domains, technical leadership experience, quality-process knowledge, and strong judgment regarding whether prompts indicate defensive, educational, research, or malicious intent.

Leads cybersecurity domain experts across reverse engineering, malware research, and vulnerability research. Manages workload allocation, quality metrics, annotation guidelines, escalated edge cases, and cross-domain reviews for AI safety prompts. Performs hands-on annotation and review when needed, coaches technical specialists, and reports team performance. Requires deep expertise in at least one security research area, working knowledge across all domains, leadership experience, and strong judgment distinguishing defensive or educational intent from attack preparation.

Lead frontier AI safety research spanning responsible AI, trust and safety, model evaluation, adversarial robustness, and agentic safety. Design benchmarks, evaluation frameworks, red-teaming methods, and safety research for text, image, multimodal, and autonomous systems. Publish research, collaborate with leading AI labs and academic institutions, advise enterprise clients, translate findings into safety products, and mentor researchers and engineers.

Conduct hands-on research on frontier AI models, reasoning, agents, and multimodal systems. Design benchmarks, evaluation frameworks, and experimental pipelines; rapidly execute research cycles from hypotheses to prototypes and validated findings. Identify model limitations and research opportunities, translate customer challenges into measurable outcomes, and collaborate with engineering and product teams. The role emphasizes LLMs, foundation models, AI security, benchmarking, and evaluation.

Conduct hands-on research on frontier AI models, reasoning, agents, and multimodal systems. Design benchmarks, evaluation frameworks, and experimental pipelines; rapidly develop prototypes, run experiments, identify model limitations, and deliver actionable findings. Collaborate with customers, engineering, and product teams to address complex research challenges. Preferred expertise includes AI cybersecurity, autonomous security agents, vulnerability research, LLM red teaming, adversarial machine learning, prompt injection, and agentic AI security.

Conduct hands-on research into frontier AI models, reasoning, agents, and multimodal systems. Design benchmarks, evaluation frameworks, and experimental pipelines; rapidly develop prototypes, run experiments, and deliver validated findings. Identify model limitations and research opportunities, translate customer challenges into measurable outcomes, and collaborate with engineering and product teams. The role requires advanced Python and PyTorch skills, strong AI research credentials, and expertise in LLMs, foundation models, benchmarking, and evaluation.

Lead and scale AI security research, red teaming, adversarial research, and guardrails engineering. Define strategy and roadmaps; direct testing of LLMs, GenAI applications, and agentic systems; research prompt injection, jailbreaks, tool abuse, and data leakage; build automated red-teaming platforms; operationalize OWASP, MITRE ATLAS, and NIST frameworks; and partner with product and engineering teams on secure AI deployments.

Lead strategy, roadmap, and execution for AI security attack and defense engines. Own adversarial testing, guardrails, evaluation metrics, benchmark datasets, data strategy, and the transition from research to production. Partner with data science, engineering, security, design, sales, and enterprise customers to define measurable quality standards, prioritize model capabilities, and translate customer risks into product requirements.

Own top-of-funnel customer acquisition for GenAI safety services through inbound and outbound prospecting. Generate qualified meetings via calls, emails, events, webinars, and campaigns; develop outreach sequences; educate prospects; map frontier AI labs, GenAI companies, and hyperscalers; and collaborate with Account Executives on lead handoffs and sales processes.

Leads end-to-end delivery, quality, scaling, and operational performance of RL Gym programs for strategic AI partners. Translates research and model-development needs into program requirements, workflows, metrics, and scalable processes. Partners with frontier-model teams, Sales, Product, Research, Delivery, and Operations to scope, launch, iterate, and expand programs. Builds feedback loops, quality standards, and the program management and operations function while managing capacity, timelines, cost, throughput, and partner satisfaction.