Applied AI Engineer

TechShack; München; Vor Ort; EUR 70.000 - 110.000; Vollzeit

Benefits dieser Stelle

Visa sponsorship

Relocation support

Learning budget

Equipment provided

Zusammenfassung

TechShack in Munich is seeking an Applied AI Engineer to take full ownership of AI features in production. You will ship LLM systems, build evaluation layers, and optimize performance for real customers.

Join a scale-up with weekly releases, guardrails, and a learning budget. Visa sponsorship and relocation support are offered to help you join the team in Munich.

Qualifikationen

• 3–5 years building software with LLM systems in production.
• Strong Python, APIs, queues, and databases.
• German at native or C2 level and fluent English.

Aufgaben

• Own AI features end to end and ship LLM systems.
• Build evaluation layers with datasets and CI tests.
• Improve accuracy via error analysis, context, guardrails and fallbacks.
• Instrument latency, cost per call and quality drift.
• Collaborate with product to decide allowed AI behavior.

Kenntnisse

Python

LLM systems

APIs

Databases

German (C2)

English (fluent)

Debugging complex AI systems

Jobbeschreibung

I'm hiring an Applied AI Engineer for a Munich based scale-up. The product is live, customers are paying, and the AI features now drive most of the usage. They were built quickly and need to hold up at significantly higher volume.

The role

You own AI features end to end.

• Ship LLM systems into production: retrieval, tool use, multi-step agents, structured extraction.
• Build the evaluation layer. Golden datasets, regression tests in CI, offline scoring.
• Move features from 80% accuracy to 95%+ through error analysis, better context, guardrails and fallbacks.
• Instrument what you build. Traces, latency, cost per call, quality drift.
• Work with product early to decide what the AI should and shouldn't do.

You’ll need

• 3 to 5 years building software, with LLM systems live in production.
• Strong Python. Maintainable code, solid with APIs, queues and databases.
• A clear view on how to evaluate model output.
• Ability to debug across prompt, model, retrieval, data and infrastructure.
• German at native or C2 level. This is a hard requirement, the role is customer-facing and internal work is conducted in German.
• Fluent English alongside it.

Bonus:

• agent frameworks, fine-tuning, AWS/GCP/Azure, or a backend background moved into AI.

Why this one

• Full ownership of an area rather than a slice of someone else's roadmap.
• Tune against real production traffic and real failure cases.
• Weekly releases. Feature flags and fast rollbacks.
• Visa sponsorship and relocation support available.
• Learning budget and your choice of equipment.

Process

Four steps, completed within three weeks.

• 30 minutes with the hiring manager.
• Scoped practical task, paid if it runs over two hours.
• Team and founder conversation, then a decision.