AI Agents & Workflows
BerlinFounding team role
AI Engineer – Applied LLMs, Workflows & Evals
Build the brains of Delvo’s procurement intelligence — reliable LLM workflows and agent systems with strong evaluation at their core.
About the role
Delvo.ai helps enterprises make confident, faster sourcing decisions. Our agentic workflows combine supplier data, price benchmarks, and risk signals with human judgment.
You will design the systems that make this reliable — building workflows, guardrails, and evaluation loops that improve continuously.
What you’ll do
- Build LLM workflows for retrieval, tool use, and structured outputs.
- Orchestrate agents with tools, control flows, retries, and safety checks.
- Create evals & monitoring (goldens, online metrics, tracing, regressions).
- Integrate with ERPs and data sources with strong observability.
- Optimise cost & latency via caching, streaming, batching, and model choice.
- Collaborate with design and forward‑deployed engineers.
Your qualifications
- TypeScript & Python across backend/product‑adjacent work (Next.js, workers, APIs).
- Hands‑on LLMs: RAG, function calling, tools, structured parsing, guardrails, streaming.
- Modern AI stack: Vercel AI SDK, OpenAI/Azure, embeddings, vector stores, Langfuse.
- Evals & quality: define tasks, gold data, and success criteria; prevent regressions.
- Bonus: data pipelines, ERP integrations, procurement domain.
What you’ll get
- Competitive salary + equity; real ownership.
- Impact & clarity; small team and visible outcomes.
- Deep technical growth in agentic systems and reliability.
- Pragmatic environment focused on simplicity and quality.
Ready to build reliable AI systems?
You like measurable progress: tracing, evals, and clean abstractions. You ship, learn, and improve systems week over week.
Apply for the role
Share your details and we’ll reach out personally. Every application is reviewed by our founding team.