Role Overview

PlexTrac is a cybersecurity SaaS platform helping security teams streamline reporting, exposure management, and remediation workflows. We are looking for an AI Research Engineer - Applied AI to build and ship the AI systems at the core of our security product. You will work across the full model lifecycle — from data pipelines and model training to deployment and production monitoring. You'll be at the forefront of our Agentic AI Offensive Security & Exposure Management Platform.

Responsibilities

Build, train, and evaluate machine learning models that detect security threats and unusual system behavior
Develop and maintain production AI features: prompt orchestration, retrieval-augmented generation (RAG), model serving, and observability
Work with raw security data — logs, network traffic, event streams — to build reliable training datasets
Build and maintain automated pipelines for model performance reporting and operational workflows
Design and maintain data ingestion and transformation services used by downstream AI systems
Monitor models in production, identify performance issues, and ship fixes
Test models for accuracy, bias, and reliability before they reach production
Work closely with security analysts to understand detection requirements and translate them into model improvements
Write clean, documented code that other engineers can read and use as a basis for implementation
Contribute to engineering standards for how the team develops and deploys models
Designing distributed training environments, optimizing computational efficiency, and managing GPU clusters.
Fine-tuning & Evaluation - Working with large language models (LLMs) and deep learning models using techniques like Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF).
Model Safety & Alignment - Testing for vulnerabilities, mitigating biases, and ensuring models behave safely and predictably.

Requirements

3+ years of software engineering experience with a focus on machine learning in production environments
Hands-on experience building and shipping ML models — not just training, but deploying and maintaining them
Strong Python skills and working knowledge of common ML libraries (scikit-learn, PyTorch, or TensorFlow)
Experience working with large, messy datasets — cleaning, labeling, and structuring data for model training
Familiarity with MLOps basics: versioning, monitoring, and retraining models in production
Ability to evaluate model performance clearly and explain trade-offs to non-technical teammates
Working knowledge of backend systems and API design

Nice to Have

Experience with security data — logs, SIEM output, network traffic, or endpoint telemetry
Background in anomaly detection, classification, or NLP applied to security use cases
Hands-on experience with LLM/RAG systems — performance tuning and reliability
Exposure to compliance-sensitive environments (SOC 2, ISO 27001, FIPS, or FedRAMP)
Familiarity with responsible AI practices — bias auditing, explainability, and model documentation
Experience with Docker or Kubernetes for model deployment
Experience with cloud ML platforms (AWS SageMaker, GCP Vertex AI, or Azure ML)
Knowledge of data privacy regulations (GDPR, CCPA) and their impact on model training
Experience with Model Context Protocol (MCP) — building or integrating MCP servers and clients