Job Specifications
About the Role
We are seeking a Distinguished AI Researcher to collaborate on the strategy, vision, and execution of the AI elements in Vectra’s groundbreaking initiative to transform the Security Operations Center (SOC). This solution harnesses classical AI/ML and Agentic AI approaches to radically transform the entire SOC operational model, enabling enterprises to more intelligently process alerts, to identify and resolve real threats faster, and to identify changes to the environment that will lead to better future outcomes.
This is a hands-on leadership role, that will require you to contribute to the development and innovation of foundation models, classical ML, graph and time-series methods, and training and evaluation techniques — from idea to prototype, and ultimately to enterprise scale production.
The role reports directly to the CTO.
Key Responsibilities
Collaborate with distinguished technical staff on design and implementation of AI solutions for end-to-end automation of detection, triage, investigation, and response across SOC workflows.
Work with Engineering and PM leaders to understand market and product requirements and develop AI strategies that enable rapid development while maximizing performance.
Rapidly adopt novel methods and architectures when advantageous, while vocally advocating traditional techniques when they are more well suited. In particular, bring a deep knowledge of how modern GenAI systems operate and where they are best used and not used.
Research and innovate state-of-the-art approaches in AI/ML/GenAI to solve the many complex, novel and evolving challenges in the domain of cybersecurity.
Publish internally/externally; work with team members to contribute to top-tier conferences and journals while also contributing to non-technical communications.
Qualifications
Advanced degree in CS/EE/Math/Physics (PhD preferred) or equivalent research/industry track record.
10+ years hands-on experience with ML/AI and 5+ years building and shipping ML systems with direct production impact.
Knowledge and expertise in the following areas: “Traditional” ML: Deep Learning, Reinforcement Learning, anomaly detection, Bayesian inference, gradient boosting, sequence models, HMMs, etc.
LLMs & GenAI: tokenizer/training pipeline design; SFT/DPO/GRPO; retrieval design (chunking, hybrid search, routing); evaluation methods (faithfulness, grounding, LLM-as-a-judge); agents (tool schemas, memory, planning).
Graph ML: entity/relationship graphs, link prediction, GNNs (e.g., variational GAE), dynamic graph learning.
Adversarial/robust ML: evasion/poisoning defenses, model red-teaming, safety/guardrails.
Experienced in python and ML frameworks (e.g., PyTorch, TensorFlow) and LLM frameworks and protocols (e.g., LangChain, Strands, MCP/A2A, etc.)
Record of publication in top-tier AI (or related) conferences and journals such as NeurIPS, ICLR, ICML, ACL, IJCAI, AAAI, etc.
About the Company
Vectra AI, Inc. is the cybersecurity AI company that protects modern networks from modern attacks. When modern cyber attackers bypass existing controls, evade detection and gain access to customers' data center, campus, remote work, identity, cloud, and IoT/OT environments, the Vectra AI Platform sees their every move, connects the dots in real-time, and stops them from becoming breaches. With 35 patents in AI security and the most vendor references in MITRE D3FEND, organizations worldwide rely on Vectra AI to see and stop a...
Know more