Giskard

We started in 2021 to provide a security layer for AI agents in the software development industry.

We offer an AI Red Teaming and LLM Security Platform that automates vulnerability scanning for Large Language Models (LLMs), from development through continuous post-deployment monitoring. Our platform helps enterprises deploy critical Generative AI systems, ensuring their AI agents are defended against threats. Our customers include AXA, BNP Paribas, Google DeepMind, Michelin, Decathlon, and Société Générale.

Our platform detects vulnerabilities like sycophancy attacks, prompt injection, data disclosure, and inappropriate content. It also addresses AI quality issues such as inappropriate denials, contradictions, hallucinations, and omissions. The Red Teaming engine continuously generates attack scenarios to test for both security and quality vulnerabilities.

Our infrastructure is sovereign and secure. We offer data residency and isolation options in the EU or US, granular access and IP controls (including Role-Based Access Control (RBAC) and audit trails), and compliance with GDPR, SOC 2 Type II, and HIPAA through end-to-end encryption. We support conversational AI agents in text-to-text mode, functioning as a black-box testing tool accessible via an API endpoint. Our testing methods use internal knowledge (like RAG knowledge bases), security vulnerability taxonomies, external resources, and internal prompt templates.

We offer a Python SDK and a web UI for testing, along with visual Human-in-the-Loop dashboards for team collaboration. We provide both an enterprise-tier Giskard Hub and an open-source version, and our solutions can be installed in on-premise environments for mission-critical applications.

Our mission is to secure conversational AI. We aim to provide the security layer that allows AI agents to run safely. We focus on preventing AI failures rather than reacting to them, enabling continuous testing of LLM agents before and after deployment.

Giskard has 11-50 employees and was co-founded by Alex Combessie and Jean-Marie John-Mathews. The team includes a Chief Technology Officer, Software Engineering Lead, Head of Sales & Partnerships, Machine Learning Engineers, AI Researchers, and Enterprise Account Executives.

It promotes a remote-friendly work environment that supports work-life balance. Its culture emphasizes collaboration across business, engineering, and security teams. They use shared tools like Human-in-the-Loop dashboards to create a common language for AI quality and security. The company focuses on growth and building repeatable enterprise sales processes in regulated sectors.

We offer a competitive compensation package that includes a base salary (ranging from 45-55 K€ or 60-75 K€ per year depending on the role), a variable salary (up to 40 K€ per year) or commission (10% on new ARR for enterprise roles), and equity (0.05% or 0.1%).

Employees benefit from a remote-flex or remote-friendly work setup, comprehensive health insurance through Alan, and a complete remote office setup (including a MacBook, external monitor, keyboard, and mouse). Our flexible work environment supports work-life balance.