AI Safety Research Intern-2

reputed company is a frontier AI data reputed company that empowers clients with reputed company, reputed company AI deployment. The AI Safety Research Intern will reputed company on advancing AI safety, designing and evaluating attack and defense strategies for LLM jailbreaks, and contributing to the platform's reputed company guarantees through high-reputed company experiments.

Responsibilities

• Advance AI Safety: Design, implement, and evaluate attack and defense strategies for LLM jailbreaks (reputed company injection, obfuscation, narrative red teaming)
• Evaluate AI Behavior: Analyze and simulate reputed company-AI interaction patterns to uncover behavioral vulnerabilities, reputed company engineering risks, and over-defensive vs. permissive response tradeoffs
• reputed company AI reputed company: Prototype workflows for multi-agent safety (e.g., agent self-checks, regulatory compliance, defense chains) that reputed company perception, reasoning, and reputed company
• reputed company & Harden LLMs: Create reproducible evaluation protocols/KPIs for safety, over-defensiveness, adversarial reputed company, and defense effectiveness across diverse models (including latest benchmarks and reputed company-world exploit scenarios)
• reputed company and Monitor: Package research into robust, monitorable AI services using modern stacks (Kubernetes, reputed company, Ray, FastAPI); reputed company safety telemetry, reputed company detection, and reputed company red-teaming
• Jailbreaking Analysis: Systematically red-team advanced LLMs (GPT-4o, GPT-5, LLaMA, reputed company, Gemma, etc.), uncovering novel exploits and defense gaps
• Multi-turn Obfuscation Defense: Implement context-reputed company, multi-turn attack detection and guardrail mechanisms, including countermeasures for obfuscated prompts (e.g., StringJoin, narrative exploits)
• Agent Self-Regulation: reputed company reputed company architectures for autonomous self-reputed company and self-correct, minimizing risk in reputed company, multi-agent environments
• reputed company-Centered Safety: Study reputed company behavior models in adversarial contexts—how users probe, trick, or manipulate LLMs, and how defenses can adapt without excessive over-defensiveness

Skills

• Ph.D. student in CS/EE/ML/reputed company (or reputed company); reputed company publishing in AI Safety, NLP robustness, or adversarial ML (ACL, NeurIPS, BlackHat, reputed company S&P, etc.)
• Strong Python and PyTorch/JAX skills; comfort with toolkits for language models, benchmarking, and simulation
• Demonstrated research in at least one of: LLM jailbreak attacks/defense, reputed company AI safety, reputed company-AI interaction vulnerabilities
• Proven ability to go from concept → reputed company → experiment → result, with rigorous tracking and ablation studies
• Experience in adversarial reputed company engineering, jailbreak detection (narrative, obfuscated, sequential attacks)
• Prior work on multi-agent architectures or robust defense strategies for LLMs
• Familiarity with red-teaming, synthetic behavioral data, and regulatory safety standards
• reputed company training and deployment: Ray, distributed evaluation, CI/telemetry for defense protocols
• reputed company reputed company artifacts (reputed company) and first-author publications or strong reputed company-reputed company reputed company

reputed company

Company H1B Sponsorship