СШАStaff Machine Learning Engineer, AI Security
1 мес. назад
230k–322k USD / годСШАLeadRemote
machine learningmodel trainingmodel optimization
Strategic and hands-on role for developing and optimizing machine learning models for AI security at .
Условия
- Comprehensive Healthcare Benefits and Income Replacement Programs
- 401k with Employer Match
- Global Benefit programs that fit your lifestyle, from workspace to professional development to caregiving support
- Family Planning Support
- Gender-Affirming Care
- Mental Health & Coaching Benefits
- Flexible Vacation & Paid Volunteer Time Off
- Generous Paid Parental Leave
- #LI-Remote
- This job posting may span more than one career level.
- In addition to base salary, this job is eligible to receive equity in the form of restricted stock units, and depending on the position offered, it may also be eligible to receive a commission. Additionally, offers a wide range of benefits to U.S.-based employees, including medical, dental, and vision insurance, 401(k) program with employer match, generous time off for vacation, and parental leave. To learn more, please visit.
- To provide greater transparency to candidates, we share base salary ranges for all US-based job postings regardless of state. We set standard base pay ranges for all roles based on function, level, and country location, benchmarked against similar stage growth companies. Final offer amounts are determined by multiple factors including, skills, depth of work experience and relevant licenses/credentials, and may vary from the amounts listed below.
- The base salary range for this position is:
- $230,000 — $322,000 USD
- In select roles and locations, the interviews will be recorded, transcribed and summarized by artificial intelligence (AI). You will have the opportunity to opt out of recording, transcription and summarization prior to any scheduled interviews.
- During the interview, we will collect the following categories of personal information: Identifiers, Professional and Employment-Related Information, Sensory Information (audio/video recording), and any other categories of personal information you choose to share with us. We will use this information to evaluate your application for employment or an independent contractor role, as applicable. We will not sell your personal information or disclose it to any third party for their marketing purposes. We will delete any recording of your interview promptly after making a hiring decision. For more information about how we will handle your personal information, including our retention of it, please refer to our Candidate Privacy Policy for Potential Employees and Contractors .
Другое
- is a community of communities. It’s built on shared interests, passion, and trust, and is home to the most open and authentic conversations on the internet. Every day, users submit, vote, and comment on the topics they care most about. With 100,000+ active communities and approximately 130 million daily active unique visitors, is one of the internet’s largest sources of information. For more information, visit.
- The AI Security team within ’s Security Platform Engineering organization builds security into ’s products, engineering systems, AI platforms, and operational infrastructure so the secure path is the easiest path for both people and agents. A core part of this work is developing practical, high-quality machine learning systems that detect and prevent risks such as prompt injection, jailbreaks, sensitive data exposure, and unsafe or unauthorized AI behavior. Building on ’s centralized LLM Guardrails Platform, which provides a shared security and safety boundary across major products, we are expanding ML-powered protections as ’s AI systems and threat landscape evolve.
- We are looking for a founding Staff Machine Learning Engineer to lead the development, training, and optimization of models for AI security at . This is a strategic and hands-on individual contributor role, combining deep ownership of model architecture, training data, and experimentation with technical leadership across teams. You will help deliver stronger AI protections while preserving a high-quality product experience.
- Select, adapt, fine-tune, evaluate, and deploy pretrained models and lightweight classifiers for -specific security problems.
- Build reproducible training and evaluation pipelines on ’s ML platform, partnering with platform engineers to improve inference performance, resource efficiency, and operational reliability.
- Set the technical vision and multi-quarter modeling roadmap, partnering with cross-functional teams to gather requirements, define model architectures, and iterate on model development.
- Conduct model evaluations and performance analysis to improve accuracy and adversarial robustness, and define launch criteria that balance false positives, latency, throughput, reliability, and cost.
- Own training-data quality and the production model lifecycle, using monitoring, incident findings, and red-team feedback to guide dataset improvements, retraining, and safe rollout or rollback.
- Establish best practices for responsible ML development and deployment, including reproducible experiments, testing, model and data lineage, and privacy-aware data use.
- Stay current with research in NLP, large language models, and relevant multimodal techniques, translating promising advances into measurable model improvements.
- Mentor engineers and lead technical discussions and reviews, shaping the team’s long-term ML capabilities and AI security direction.
- Who You Might Be
- 8+ years of experience developing machine learning models, with substantial hands-on model training experience, demonstrated production impact, and a record of leading complex initiatives across teams.
- Strong background in Python programming, software engineering, and deep learning frameworks and libraries such as TensorFlow, PyTorch, or Hugging Face Transformers.
- Deep understanding of neural network architectures and optimization, with proficiency in data preprocessing, tokenization, embeddings, language modeling, and model calibration.
- Expertise in scalable data pipelines and distributed training frameworks such as Ray Train or PyTorch Distributed, with a strong understanding of hardware and system tradeoffs.
- Demonstrated rigor in experimental design and model evaluation, including representative holdouts, ablation studies, adversarial tests, and error analysis to diagnose training issues, bias, and generalization gaps.
- Excellent written and verbal communication, with the ability to explain model behavior, security risk, uncertainty, and tradeoffs to technical and non-technical partners.
- Preferred Qualifications
- Experience applying ML to security, trust and safety, fraud, privacy, or related adversarial domains.
- Experience with adversarial training, model distillation, active learning, or synthetic-data generation to improve model quality, robustness, and training efficiency.