Principal AI Engineer, Product Cybersecurity

About TTS-US:

Founded in 2011, Toyota Tsusho Systems US, Inc. (TTS-US) is a Toyota group company, that develops IT solutions wherever global businesses operate. Transforming into a technology and mobility company, TTS-US with it’s 8 TTS affiliates worldwide is establishing a secure and resilient Toyota global value chain. The creative capacity to forge such limitless business opportunities is one of the strengths of Toyota Tsusho Systems.

Summary:

The Product Cybersecurity Group is building an AI harness that brings Generative and agentic AI capabilities to Product Cybersecurity engineering. The Principal AI Security Software Engineer works hands-on with the team, building the prompts, agents, tools, connectors, and workflows that make AI perform real Product Cybersecurity work across PCG workflows and customer environments.

This is a hands-on engineering role. The Principal AI Security Software Engineer spends most of their time writing, testing, and shipping code alongside the team, rather than producing architecture documents for others to implement. This role does not train foundation models. Instead, the engineer takes existing models and makes them perform useful, reliable, measurable work by building the full path from prototype to production, including the data access, orchestration, evaluation, and observability that make AI features trustworthy enough for use in security-sensitive environments.

The Principal AI Security Software Engineer brings a software developer’s mindset to AI, including version control, code review, testing, debugging, and maintainable design. The engineer works as part of the team, partnering with engineers, turning security tasks into working AI-assisted workflows, and helping team members learn to build and extend those workflows themselves.

Essential Functions:

· Build (the majority of the role)

o Write and ship the components that make AI do the work — prompts, agents, skills, tools, MCP servers, connectors, scripts, and orchestration logic — assembled into end-to-end LLM applications such as retrieval pipelines, agent workflows, and structured-output workflows

o Turn the team’s manual security tasks into working AI-assisted workflows, such as engineering analysis, knowledge retrieval, documentation, requirements analysis, code/design review, test generation, and security-assessment support

o Securely connect the harness to approved customer documentation, knowledge sources, repositories, APIs, identity systems, and downstream tools so AI operates with relevant context

o Build a reusable library of components so each new use case is faster to deliver than the last

· Shape the technical approach

o Make and document practical design decisions for the harness — model integration, context management, tool integration — so the team can build on them

o Contribute design and code reviews that keep components maintainable, secure, and reusable

o Recommend when to build, reuse, or stop an approach based on evidence from working implementations

· Measure

o Build evaluations and quality gates for the workflows you ship — covering output quality, groundedness, hallucination rate, reliability, security, usefulness, latency, and cost — so releases rest on evidence rather than demo impressions

o Analyze production transcripts and user feedback to drive iterative improvement

· Operate

o Version, test, deploy, monitor, and maintain prompts, context, and components with the same discipline as any other code

o Design for reliability — provider or model fallback, graceful degradation, timeouts, and retries — and treat inference cost and latency as first-class constraints

o Implement security practices — access controls, data handling, traceability, guardrails — aligned with internal and customer requirements, and release changes through existing change-management and security review processes

· Work with the Team

o Pair with team members and customer engineers to understand their work, identify the highest-value tasks for AI, and build solutions together

o Help define what AI should and should not be used for, explaining technical constraints in terms non-specialists can act on

o Teach by building: write clear examples, document how components work, and help engineers create and extend their own AI workflows

Competencies:

· Advanced software engineering and application development expertise.

· Strong knowledge of Generative AI, agentic AI systems, prompt engineering, context engineering, retrieval-augmented generation (RAG), and LLM application development.

· Ability to evaluate, troubleshoot, and improve complex AI-enabled workflows using measurable performance criteria.

· Strong written and verbal communication skills with the ability to explain technical concepts, risks, and tradeoffs to technical and non-technical audiences.

· Ability to collaborate effectively across cross-functional teams and mentor engineers through hands-on development and knowledge sharing.

· Demonstrated critical thinking, innovation, sound judgment, and the ability to solve complex technical problems in security-sensitive environments.

Requirements

· Bachelor’s degree in Computer Science, Engineering, or a related field, or equivalent practical experience.

· 8+ years of professional software development experience, with a track record of personally writing, shipping, and maintaining production software.

· Solid software engineering fundamentals: version control, code review, automated testing, debugging, and maintainable design.

· Hands-on experience building LLM-powered applications or AI-assisted workflows, including prompting and context engineering, retrieval (RAG), agent and tool-use frameworks, and evaluation; production experience strongly preferred.

· Practical experience with model APIs from at least one major provider (Anthropic, OpenAI, Google, or an open-weight equivalent).

· Cloud experience (AWS, Azure, or GCP), containerization, and CI/CD.

· Experience designing secure systems: access controls, data handling, traceability, and guardrails for sensitive data.

· Ability to reason rigorously about non-deterministic systems and understand why a passing demo is not equivalent to a passing test.

· Strong written and verbal communication; able to explain tradeoffs to engineers and executives alike.

Preferred Qualifications:

· Experience building agentic AI systems: tool usage, workflow orchestration, multi-step autonomous workflows, and human-in-the-loop patterns.

· Experience with MCP, agent orchestration frameworks, and observability/tracing for LLM applications.

· Background in Product Cybersecurity, application security, or security-assessment workflows.

· Experience defining evaluation frameworks for AI output quality, reliability, and safety.

· Familiarity with MLOps/LLMOps practices for versioning, deploying, monitoring, and maintaining AI systems.

· Familiarity with AI governance and risk frameworks (NIST AI RMF, ISO/IEC 42001, EU AI Act) and secure handling of sensitive data in AI systems.

· Experience running technical workshops, pairing, or mentoring engineers new to AI.

Apply for this job

Source: Toyota Tsusho Systems job board (Workable (all companies)).

Report this job

Wrongly listed? Tell us why.