SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Tempo is seeking a Senior AI/ML Engineer to build production AI systems at the intersection of LLMs, real-time signal processing, and enterprise decision-making. With 30,000+ customers including a third of Fortune 500 companies, Tempo is the #1 time management add-on for Jira and a trusted name in the Atlassian ecosystem.
This role is not isolated research or a typical AI team position. You'll work embedded within a product team, shipping production systems to enterprise customers alongside domain engineers, product managers, and designers.
Key responsibilities include:
**Signal Detection & Anomaly Detection**: Build statistical and ML-based detectors that identify meaningful patterns in portfolio signals—velocity changes, capacity saturation, dependency risks—from CDC event streams and external tool integrations. Move beyond simple threshold alerts to intelligent pattern recognition.
**Insight Synthesis Engine**: Design an LLM-powered correlation engine that transforms raw signals into actionable insights with root causes, confidence scores, and evidence chains. Provide not just problem identification but context and recommended actions.
**Planning Rules Compiler**: Create a translation layer between natural language planning rules (written by portfolio managers) and structured parameters for the Monte Carlo scheduling engine. Design how LLM interpretation and deterministic computation communicate reliably.
**Evaluation & Testing Frameworks**: Build pipelines ensuring AI outputs are reliable, consistent, and improving over time. Implement regression suites for prompt changes, A/B testing for model updates, and confidence calibration—moving beyond subjective assessment to enterprise-grade validation.
**MCP Tool Definitions**: Define LLM-ready tool specs for domain capabilities (Item Store queries, capacity lookups, scenario simulations) that Tempo AI can discover and invoke at runtime within an existing hub-and-spoke MCP architecture.
You'll initially pair with external AI partners to absorb their work and establish patterns for the broader engineering team. Over time, you become the internal anchor for AI engineering—the person the team turns to for prompt design, evaluation strategy, model selection, and agent architecture decisions.
**Requirements:**
Must Have:
- Track record of shipping LLM-powered features or products to real users (prototypes don't count)
- Hands-on experience orchestrating agents with multi-step reasoning, tool use, and autonomous action with guardrails (LangChain, LlamaIndex, CrewAI, AutoGen, or equivalent)
- Deep LLM engineering fundamentals: prompt engineering, RAG architectures, function calling/tool use, context management, evaluation-driven development
- Production-quality engineering practices: testing, code review, CI/CD, observability; systems that run reliably in production
- Experience with event-driven or streaming data systems (CDC events, real-time pipelines)
- 5+ years in software engineering, with 3+ years focused on AI/ML in production systems
- Ability to work embedded in a product team, collaborating daily with domain engineers, product managers, and designers
Strongly Preferred:
- AWS Bedrock, Azure OpenAI, or GCP Vertex AI (Tempo currently uses Bedrock with Claude)
- MCP (Model Context Protocol) or similar tool-use/agentic frameworks
- Anomaly detection, time-series analysis, or statistical signal processing
- Building confidence scoring or calibration systems for AI outputs
- Track record of absorbing work from external partners and improving inherited architectures
- Proficiency in Kotlin or TypeScript alongside Python (product platform is Kotlin/TypeScript; AI platform is Python)
Nice to Have:
- Monte Carlo simulation, optimization algorithms, or scheduling systems
- Background in portfolio management, project management, or resource planning
- Enterprise SaaS experience (multi-tenancy, compliance, audit trails)
- Open-source AI/ML contributions
Mindset:
- Use AI tools heavily in daily work for meaningful productivity gains
- Strong opinions (loosely held) on trustworthy vs. impressive-but-fragile AI products
- Energized by fast-moving space; track model releases, emerging frameworks, new patterns
- Comfortable writing Architecture Decision Records and prompts with equal fluency
