Software Engineer - Machine Learning III

Ony apply if candidate is Local to MountainView CA and only on W2!!

Machine Learning Engineer - Prompt Safety & Agent Security | Contract

Protect intelligent agentic AI systems for a leading consumer electronics innovator by designing and training advanced prompt injection detection models, building hybrid on-device and cloud safety guardrails, and applying cutting-edge post-training techniques to defend against adversarial attacks.

About the Role: We're seeking an experienced Machine Learning Engineer specialized in AI safety to lead prompt safety and agent security for our client in the consumer electronics and intelligent devices sector. This contract role offers the opportunity to design production-grade safety models that protect agentic AI systems across mobile, cloud, and XR/AR platforms — building the guardrails that enable billions of users to interact safely with intelligent assistants.

Key Responsibilities:

• Design and train prompt injection detection models and prompt safety classifiers for both inputs and outputs of agentic AI systems
• Build hybrid deployment pipelines splitting safety inference between on-device (phone, XR/AR) and cloud optimizing for latency, privacy, and detection coverage
• Apply post-training techniques (RLHF, DPO, RLAIF, reward modeling) to optimize guardrail model performance, calibration, and robustness against adaptive adversaries
• Curate and generate adversarial training data including prompt injections, jailbreaks, tool-use exploits, and unsafe-output cases from red-teaming and production signals
• Build evaluation harnesses measuring attack success rate, false-positive rate, latency, and on-device footprint across model iterations and threat categories
• Integrate safety models into mobile agents, XR/AR assistants, and cloud agentic workflows; close the loop from production incidents into training data
• Design reward models and preference data curation strategies for post-training safety alignment
• Conduct adversarial robustness testing and red-teaming to identify model vulnerabilities and improve defenses
• Collaborate with security researchers, modeling teams, and product engineers on safety strategy and threat modeling
• Document safety methods and contribute to patents and academic publications where appropriate

Requirements:

• M.S. or Ph.D. in Computer Science, Machine Learning, Electrical Engineering, or related field (or B.S. with equivalent industry experience)
• 3 years industry experience in ML engineering or applied AI research with demonstrated ownership of production ML systems
• 2 years industry experience in software engineering
• Expert proficiency in Python and PyTorch (or JAX/TensorFlow) with solid software engineering fundamentals
• Hands-on experience post-training LLMs with RLHF, DPO, RLAIF, or reward modeling
• Demonstrated experience training and deploying classifier/guardrail models for safety, content moderation, or adversarial robustness
• Deep understanding of prompt injection, jailbreak, and agentic AI threat models
• Experience with distributed training frameworks (DeepSpeed, FSDP, Accelerate)
• Strong version control, testing, and reproducible experimentation practices
• Reward design, preference data curation, and training stability expertise

Preferred Qualifications:

• Experience building safety/moderation systems for agentic AI (tool-use guardrails, indirect prompt injection defense, output filtering)
• Red-teaming, adversarial data generation, or automated attack pipeline experience (GCG, PAIR, generator-critic frameworks)
• On-device/edge ML deployment expertise (ExecuTorch, Core ML, TFLite, MLC-LLM, NPU toolchains)
• Model compression experience (quantization, distillation, pruning) for safety models
• Telemetry, logging, or user-facing data systems on mobile/XR/AR platforms
• Privacy-preserving data handling (anonymization, on-device processing, federated approaches)
• Publications at top-tier ML/NLP/security venues (NeurIPS, ICML, ICLR, ACL, EMNLP, USENIX Security, IEEE S&P)
• Patents or open-source contributions in AI safety, alignment, or security
• Experience optimizing safety models for hybrid on-device/cloud deployment

Contract Details: Contract Position (Extension Possible) | Consumer Electronics/Intelligent Devices | Reports to ML Research Leadership

Next Steps: Submit your resume highlighting your post-training experience (RLHF/DPO), guardrail model development, and adversarial robustness work. Include any publications, patents, or open-source contributions in AI safety.

#MachineLearning #AISafety #PromptSafety #RLHF #DPO #AgentSecurity #LLMSafety #AdvancedML #ContractWork #TechJobs