Senior Applied AI Engineer

Axiomatic AI is building innovative AI systems that combine deep learning with formal logic to support scientific and engineering workflows. The Senior Applied AI Engineer will bridge the gap between AI research and production software, focusing on developing AI features and collaborating with stakeholders to enhance workflows.

Responsibilities:

• Partner with users and internal stakeholders to identify high-impact workflows where AI can help
• Design and implement LLM-powered product features (agents, tools, prompting strategies) using frameworks like PydanticAI (or similar)
• Make principled trade-offs between models and approaches (quality, latency, cost, privacy, reliability)
• Enable AI developers to deploy their work reproducibly and safely
• Establish engineering standards for applied AI development: testing, reviews, maintainability, and operational readiness
• Collaborate with backend engineers to integrate AI capabilities into the product
• Mentor junior AI developers and supervise PRs to ensure high quality and consistent patterns
• Own applied AI features end-to-end: discovery → design → implementation → rollout → iteration
• Translate user feedback into clear technical requirements and pragmatic delivery plans
• Build LLM workflows such as tool-calling agents, structured output pipelines, retrieval/tool integrations, and safe prompting strategies
• Iterate quickly while keeping production quality (readability, maintainability, debuggability)
• Select and evaluate LLMs (OpenAI/Anthropic/others) based on real constraints: quality, cost, latency, context limits, and reliability
• Develop prompt patterns and guardrails (structured prompts, schemas, constraints, fallbacks)
• Design and run lightweight evaluations to prevent regressions (golden datasets, acceptance criteria, failure-mode testing)
• Document model decisions and trade-offs in a way that enables other engineers to execute confidently
• Write production-grade code: clear abstractions, solid API boundaries, strong typing where appropriate, and consistent error handling
• Define and enforce testing practices for applied AI (unit tests, integration tests, golden/regression tests)
• Implement instrumentation appropriate for debugging and iteration (basic logging/tracing/metrics for AI features)
• Ensure reliability and security basics: rate limiting where needed, safe input handling, prompt-injection awareness, and sensible defaults
• Work with AI Developers to productionize their experiments regarding improving user workflows
• Define workflows: notebook/test repository → PR → staging → production
• Document AI infrastructure and best practices
• Review code and mentor AI developers on software practices