AI Engineer
We are looking for engineers excited about building long-lived AI systems, not chatbots. As a Founding AI Systems Engineer, you own the AI core of the platform: agents, tools, ontology generation, memory, retrieval, and evaluation. This is not a model-training role. The work is agent architectures, semantic retrieval, knowledge representation, enterprise data systems, decision intelligence, and autonomous learning from operational exhaust: the workflow steps, approvals, log entries, and data changes an enterprise produces as it runs.
Above all, you build in an LLM-first, reasoning-first way. Nearly everything you ship should make the platform a little more self-improving, the way Anthropic let Claude Code help write Claude Code.
About the Role
LLM orchestration and tool calling: the frameworks StarLifter’s agents use to reason over enterprise context and act.
Agent frameworks: agentic systems that observe signals, reason about decision patterns, and recommend or automate governed actions.
Ontology generation: pipelines that induce a machine-readable model of each enterprise (its products, customers, orders, rules, and policies) from its own systems and knowledge bases.
Embedding pipelines and knowledge graph integration: the semantic retrieval layer that serves enterprise behavior back to agents.
Operational memory: the durable record of events, decisions, actions, and outcomes that lets the platform learn from every decision.
Evaluation harnesses: the systems that score confidence, benchmark outcomes, and prove decision quality improves over time.
Requirements
Real depth in modern AI tooling: LLM orchestration, RAG architectures, retrieval systems, and knowledge graphs in production. This field is young (these tools are only a couple of years old), so we care more about how far you’ve pushed them, and how fast you learn, than about years on a résumé.
Hands-on with agent frameworks and orchestration tooling such as LangGraph and DSPy.
Strong engineering fundamentals and comfortable in a modern typed stack (we build in TypeScript + Node). You write production-quality code and think about reliability and scale by default.
Genuinely excited by enterprise ontology induction, operational memory, decision patterns, agentic systems, and learning from operational exhaust.
The kind of engineer who tests assumptions and shows up knowing more than we do about deep AI. You’ll challenge our thinking and make the platform better for it.
Intentional about your work and your career: you build meaningful, long-lived systems and stay to see them through, energized by the ownership and ambiguity of an early-stage company.
A craftsperson and a colleague: low ego, high standards, excited to build alongside a small, senior, highly collegial team.
Nice to have:
Working familiarity with core enterprise business processes (Quote-to-Cash, Procure-to-Pay, Hire-to-Retire) and ERP data, enough to understand what the platform is reasoning about.
Semantic modeling, metadata systems, or business-context modeling experience.
Deep familiarity with enterprise integrations (SAP, Oracle, ServiceNow, Salesforce, Snowflake, Databricks).
Prior founding-engineer or very-early-startup experience.