Principal AI Engineer

Spacewalk AI, USA  |  (December 2025 – Present)

Building AI-powered incident response solutions that help security teams investigate and manage complex breaches through automated timeline generation, IOC extraction, and intelligent reporting.

Staff Data Scientist

Obsidian Security, USA  |  (2025)

Led Generative AI initiatives, developing both internal analyst-facing tools and external customer-facing solutions to accelerate SaaS security investigations.

• Obsidian AI Assistant: Core developer of a governed multi-agent AI system that helps security teams triage threat detections, identify high-risk identities, and evaluate security posture—all with explainable, auditable recommendations.
• Built comprehensive evaluation framework for agent performance, tracing, and agentic execution monitoring.
• Developed Gen-AI tooling to accelerate security content creation by analysts, integrating agentic workflows, web search capabilities, and a RAG system to surface security-relevant insights about SaaS services.

Security Data Scientist

Obsidian Security, USA  |  (2022 – 2024)

Focused on developing advanced threat detection models and monitoring systems for cybersecurity solutions in SaaS environments.

• Designed and deployed multiple threat detection models for SaaS environments, targeting account compromise and behavioral anomalies.
• Built monitoring pipelines to evaluate and maintain detection model performance, ensuring long-term accuracy and minimizing alert fatigue.
• Contributed to posture management features through scalable data extraction, modeling, and orchestration workflows.

PhD Researcher

State University of New York at Binghamton, USA  |  (2018 – 2022)

Dissertation: "Empowering Artificial Intelligence for Cybersecurity Applications"

During my PhD, I focused on using AI and machine learning for improving cybersecurity. My research aimed to design real-time models for anomaly detection, improving explainable malware classification, and applying blockchain to enhance data integrity in scientific workflows.

• CFGExplainer: Developed a model that explains malware classification decisions made by Graph Neural Networks (GNNs).
• Log-Anomaly-Mask: Designed a real-time adversarial evasion attack against deep learning-based system log anomaly detection.
• RAMP: Built a real-time machine learning model for anomaly detection in streaming multivariate time series data.
• SciBlock: Utilized blockchain technology to enable tamper-proof storage and improved reproducibility in scientific workflows.

Graduate Research Assistant

State University of New York at Binghamton, USA  |  (2017 – 2018)

In this role, I conducted research in wireless networks, focusing on deep learning for wireless channel quality prediction and Markov models for routing in cache networks.

• DeepChannel: Developed an LSTM/GRU-based deep learning model for wireless channel quality prediction.
• Opportunistic Routing in Cached Wireless Networks: Designed a Markov model to analyze routing behavior in wireless cache networks.