Principal Software Engineer – Applied AI & Agentic Systems

Your mission

Join us as our Principal AI Software Engineer and help turn Matrix42's  AI Your Way  vision into secure, reliable, and useful product capabilities.

This is a principal-level, hands-on software engineering role. You will spend most of your time designing, coding, testing, and operating production software. You will also be a technical multiplier: working alongside developers when the problems are difficult, turning working experiments into reusable components, and helping product squads make sound architectural decisions.

You will connect real customer and product needs with the right combination of software, data, and AI. Sometimes that means a deterministic workflow; sometimes retrieval, tool calling, or an agentic system. Your job is to choose pragmatically, prove value early, and build a path from the first working vertical slice to a scalable product capability.

Your key responsibilities would be:

• Own selected AI capabilities end to end—from problem discovery and technical design through implementation, evaluation, release, production monitoring, and continuous improvement.
• Write and review production code across AI services, AI harnesses, APIs, connectors, background services, MCP-compatible tools, data pipelines, and the product surfaces needed to deliver a complete workflow.
• Design agentic systems using the simplest architecture that works, including retrieval and grounding, structured outputs, tool execution, state and context, model routing, approval flows, fallbacks, and graceful degradation.
• Translate each use case into concrete data and platform requirements. Work with APIs, event streams, ingestion and transformation, data quality and alignment, metadata, storage, observability, and governance so that AI results are trustworthy.
• Create results early: build focused end-to-end prototypes with small teams, validate them against real workflows, and evolve successful patterns into secure, maintainable, multi-tenant product capabilities.
• Make evaluation part of engineering. Build representative datasets, automated and human-reviewed evaluations, trace analysis, regression gates, and telemetry for quality, task completion, latency, safety, and cost.
• Engineer for enterprise trust through tenant isolation, least-privilege tool access, identity and authorization, auditability, data minimization, prompt-injection defenses, human approval for consequential actions, feature flags, and safe rollback.
• Collaborate closely with Product, Design, Architecture, Security, Support, Customer Success, and engineering teams. Turn customer problems into measurable outcomes, unblock developers, contribute reusable libraries and reference implementations, and raise applied-AI engineering practices across Matrix42.

Your profile

Must haves

• 7+ years of professional software engineering experience, or equivalent evidence of senior/principal-level impact in production product development.
• A strong record of shipping and operating customer-facing software—not only notebooks, proofs of concept, demos, or advisory work.
• Professional proficiency in Python and strong ability in at least one product engineering language such as C#, TypeScript, or Java.
• Hands-on experience building production LLM or agentic applications, including several of the following: retrieval and grounding, embeddings or hybrid search, structured outputs, tool or function calling, context and state management, model selection or routing, and human approval or escalation.
• Experience with data-intensive systems: APIs and integrations, event or streaming data, ingestion and transformation, data quality, metadata, storage, and operational observability.
• Strong system-design fundamentals for API-first, distributed, cloud-native, multi-tenant SaaS products, including identity, authorization, reliability, performance, and secure integration with enterprise systems.
• Practical experience evaluating and diagnosing AI systems through datasets, traces, qualitative review, quantitative metrics, automated tests, CI/CD, and production telemetry.
• Pragmatic architectural judgment. You know when conventional software is sufficient, when AI adds real value, and how to balance fast validation with a credible path to production.
• Clear communication and product thinking. You can work directly with developers and customers, translate ambiguous needs into a testable technical plan, explain trade-offs, and drive an end-to-end result across team boundaries.
• A degree in computer science, software engineering, AI/ML, or a related field—or equivalent practical experience.

Nice to have

• Experience with observability, telemetry, operational intelligence, or other high-volume event-data platforms.
• Knowledge of IT Service Management, Enterprise Service Management, endpoint management, knowledge management, service automation, or employee self-service.
• Experience with Microsoft Azure, Azure …