Senior Applied AI Workflow Engineer
Python · LLM Workflows · Reliable Automation
Berlin / Remote (EU preferred)
Context
Talonic builds enterprise AI data products that turn messy operational inputs (documents, systems, workflows) into structured, system-ready datasets.
This role is not about building “AI agents.” It is about building reliable AI-assisted workflows that may be agentic when necessary, but are always observable, controllable, and production-safe.
We are working with enterprise customers and are an early-stage startup with a steep growth trajectory. You can build something that matters here. Delivery credibility is non-negotiable.
What you will build
You will own workflow-level AI product surfaces such as:
- AI-assisted data consolidation across heterogeneous sources
- Schema-driven mapping and normalization workflows
- Multi-step structuring pipelines (documents → dataset → export)
- Tool-using LLM workflows where automation is phased
- Human-in-the-loop workflows with clear failure visibility
- Enterprise-grade workflow behavior, not chat outputs
The output is not “a model response.”
The output is a correct, usable dataset and a debuggable system state.
Your mandate
You are a senior IC with full ownership of applied AI workflows end-to-end.
You will:
- Design workflow architectures that survive real variance
- Implement production-grade LLM pipelines in Python
- Decide what must be deterministic vs AI-assisted
- Build explicit state transitions, retries, and safe degradation
- Create evaluation and QA mechanisms for workflow correctness
- Ship iteratively into real customer environments
- Own failures, trade-offs, and operational behavior
This role exists to reduce founder arbitration by creating owned, dependable AI workflow surfaces.
What this role is not
- Not an academic ML research role
- Not prompt experimentation without delivery
- Not a “multi-agent autonomy” playground
- Not prototyping and handing off to others
If you don’t ship production systems yourself, this is not the role.
Required profile (non-negotiable)
You likely have 7+ years experience and:
- Strong Python engineering skills
- Hands-on experience shipping LLM-powered workflows into production
- Ability to reason about state, failure modes, and system invariants
- Experience integrating AI behavior into enterprise software constraints
- Comfort owning implementation end-to-end under delivery pressure
You think in:
- Workflow correctness over model cleverness
- Explicit failure over silent success
- Systems behavior over demos
Strong signals
- Built AI-assisted automation beyond trivial prompt chains
- Experience with tool orchestration frameworks (or equivalent custom stacks)
- Experience building evaluation harnesses, monitoring, and rollback paths
- Product instinct: knows what not to automate yet
- Comfortable working close to real enterprise delivery teams
Team & structure
- You are a senior individual contributor
- You work closely with:
- CEO (product ownership)
- CTO (delivery pressure, needs load reduction)
- Data Pipeline Owner (separate role, owns extraction/pipelines)
- Senior Full-Stack Engineers (Platform & Front-End, execution-focused)
- You are not expected to manage people
Hiring philosophy
We optimize for:
- Ownership over breadth
- Data Correctness over Integrity
- Engineers who think in systems, not screens
False positives are extremely costly for us.
Compensation
- Base salary: €80,000 - €90,000 gross per year, depending on experience and demonstrated ownership
- Equity: Meaningful equity ownership in Talonic, subject to role fit and long-term commitment
- Benefits: Standard package (equipment budget, Deutschlandticket, flexible remote setup)
.png){kind=logo}
.png)
