Urgently required:: AI/ML Security Researcher II _ Redmond, WA(Onsite)@Long-Term Contract

[Vasu Naidu]

Hello ,

Greetings for the day.

I am trying to reach you for the below position .Please let me know your interest in taking the new positions or else please refer your colleagues or friends who are looking for a job change.

 

Role: AI/ML Security Researcher II 

Location: Redmond, WA(Onsite)

Duration: Long-Term Contract

Client: Microsoft

ABOUT THE ROLE

 

 

Security Graph's AI layer is responsible for transforming raw entity relationship data into intelligent, real-time security signals that protect enterprises worldwide. As an AI/ML Security Researcher II, you will design and implement machine learning models, contribute to graph-based feature engineering pipelines, and apply data science techniques to threat detection challenges within the Security Graph ecosystem.

This role is ideal for an ML practitioner with 3–6 years of experience who is eager to apply cutting-edge techniques—including graph neural networks, anomaly detection, and LLM-based security reasoning—to real-world security problems at massive scale. You will collaborate closely with senior researchers, security engineers, and product teams to build detection systems that power Microsoft Sentinel, Defender XDR, and Security Copilot.

 

KEY RESPONSIBILITIES

 

 

ML Model Development & Experimentation

• Develop, train, and evaluate machine learning models for threat detection use cases including anomalous entity behavior detection, identity risk scoring, lateral movement identification, and data exfiltration risk prediction.
• Apply graph machine learning techniques (graph neural networks, node embedding algorithms, link prediction) to Security Graph data representing users, devices, applications, and their relationships.
• Implement and benchmark model architectures—including gradient boosting (XGBoost/LightGBM), neural networks, and transformer-based models—for tabular and graph-structured security data.
• Conduct rigorous offline evaluation using precision/recall, ROC-AUC, and security-specific metrics; contribute to online A/B testing frameworks for production model validation.

Feature Engineering & Data Pipelines

• Design and implement feature extraction pipelines that convert Security Graph signals—node properties, edge attributes, traversal depth, temporal activity patterns—into ML-ready features.
• Work with large-scale security telemetry from Microsoft Entra ID, Defender products, and unified audit logs using Azure Databricks, PySpark, or Azure ML.
• Collaborate with data engineers to define schema requirements, ensure data quality, and optimize pipeline performance for the Security Graph's real-time refresh cycle.
• Explore unsupervised and semi-supervised learning techniques for detecting novel, unseen attack patterns in sparse labeled security datasets.

Security Copilot & LLM Integration

• Contribute to the development of LLM-powered security reasoning capabilities within Microsoft Security Copilot, including prompt engineering, retrieval-augmented generation (RAG) over graph data, and tool-use agent patterns.
• Evaluate LLM outputs for security accuracy, hallucination risk, and operational reliability in high-stakes security contexts.
• Develop benchmark datasets and evaluation harnesses for assessing AI agent performance on security graph traversal and threat summarization tasks.

Research Support & Knowledge Sharing

• Support senior researchers in preparing research papers, technical reports, and internal documentation on AI-driven security detection methods.
• Contribute to MLOps practices including model versioning, experiment tracking (MLflow), and deployment pipelines for production security ML models.
• Participate in internal research reviews, hackathons, and cross-team knowledge-sharing sessions within the Microsoft Security AI community.
• Stay current with advances in graph ML, LLMs, and AI security research; regularly present literature reviews and new technique evaluations to the team.

 

REQUIRED QUALIFICATIONS

 

 

• Bachelor's or Master's degree in Computer Science, Machine Learning, Data Science, Statistics, or a related quantitative field.
• 3–6 years of hands-on experience in applied machine learning, data science, or ML research, with production model development experience.
• Strong proficiency in Python and the core ML stack: PyTorch or TensorFlow, scikit-learn, pandas, NumPy.
• Experience with at least one graph ML framework (PyTorch Geometric, DGL, GraphSAGE) or graph analytics tools.
• Solid understanding of supervised, unsupervised, and semi-supervised learning techniques with practical experience applying them to real-world datasets.
• Familiarity with cloud ML platforms—Azure ML, AWS SageMaker, or GCP Vertex AI—for model training and experiment management.
• Basic understanding of cybersecurity concepts: identity and access management, threat detection, network security, or endpoint security.
• Strong analytical skills and ability to communicate ML results clearly to both technical and non-technical audiences.

 

PREFERRED QUALIFICATIONS

 

 

• Experience working with LLMs, prompt engineering, RAG architectures, or LLM-based agent frameworks (LangChain, Semantic Kernel).
• Exposure to anomaly detection, time-series analysis, or behavioral analytics in security or fraud detection contexts.
• Familiarity with Microsoft Sentinel, Defender XDR, Azure Databricks, or the Microsoft Intelligent Security Graph platform.
• Experience with MLOps tools: MLflow, Azure ML Pipelines, DVC, or Weights & Biases.
• Knowledge of responsible AI, explainable AI (XAI), or fairness in ML.
• Research publications, Kaggle competition achievements, or open-source ML contributions.

 

 

Thanks & Regards,

Vasu.K

US IT Recruiter

Email: va...@atvsllc.com

 

DISCLAIMER: This email and any files transmitted with this email are confidential and intended solely for the use of the recipient to whom they are addressed. If you are not the intended recipient you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this mail and attached file/s is strictly prohibited. You are also advised to notify the sender immediately by e-mail if you have received this e-mail by mistake and delete this e-mail from your system.

 

This e-mail message including attachment(s), if any, is believed to be free of any virus.  However, it is the responsibility of the recipient to ensure that is virus free and American Tech Vision Solutions is not responsible for any loss or damage arising in any way from its use.