SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
SandboxAQ is a high-growth AI company emerging from Alphabet Inc. (2022) that develops Large Quantitative Models (LQMs) addressing challenges in life sciences, financial services, navigation, cybersecurity, and other sectors. The company is backed by leading growth capital investors and operates as a global, multidisciplinary team.
The AI Engineer role focuses on building agentic collaboration, trust, and ROI. You will work at the intersection of AI and software engineering, contributing to three core areas:
**Security**: Agent discovery, evaluation, red teaming, monitoring, and guardrails to protect deployed agents from misuse and attacks.
**Quality**: Measuring agent ROI and suggesting improvements through classical ML tasks like exploratory data analysis, feature engineering, and model fine-tuning.
**Collaboration**: Creating frameworks that bring humans and agents together with shared context, enabling effective teamwork.
Key responsibilities include:
- Creating new agentic capabilities and frameworks for monitoring, guardrailing, and evaluating agents
- Performing classical ML tasks (EDA, feature engineering, fine-tuning) using latest frameworks
- Translating AI research into applicable features and software engineering projects with Product and Research teams
- Rapidly iterating on prototypes and features based on feedback
- Contributing production-ready, scalable code across the engineering organization
- Staying current with latest agentic developments and exploring new opportunities
This is a hands-on role where you'll prototype, break things, fix them, and harden solutions for production. You'll have freedom to explore, collaborate, publish, and deploy, shaping the field of agentic collaboration and trust. You'll work alongside ML experts, security experts, mathematicians, and software engineers.
The ideal candidate combines the curiosity of a researcher, rigor of an engineer, and creativity of a hacker. You deeply understand how models and agents work, how they fail, how they can be protected and improved. You have hands-on production experience with agents, know what works and what doesn't, and stay current with trends. You're opinionated about the future, know what you don't know, and enjoy learning to close gaps.
**Requirements:**
- MSc in Computer Science or related field with exposure to AI/ML
- Deep expertise in AI/ML, software engineering, or AI security
- Strong programming skills in Python and relevant ML/agentic frameworks
- Demonstrated ability to work cross-functionally with high degree of end-to-end ownership, delivering real-world outcomes
- Ability to pick up other programming languages and technologies as needed to ship scalable, reliable software
**Highly Desired:**
- PhD in Computer Science or related field with focus on AI/ML
- Hands-on experience with: agent orchestration, agent harnesses, agent teams, jailbreaking/prompt injection, red teaming, data poisoning or evasion attacks, secure model deployment or sandboxing, detection and monitoring for AI misuse
- Open source project contributions
- Fast-paced startup environment experience
- Cybersecurity background (a plus, not essential)
