AI / LLM Engineering (Werkstudent)

About the company

CognitX AI GmbH is developing a privacy-focused AI advisor for analytics and process automation. The product enables users to ask complex questions about their data in natural language and receive dependable answers, visual insights, and automated actions. Its architecture is designed to connect directly with enterprise data sources while running analyses through a secure compute layer. Both on-premise and cloud deployment options are treated as core capabilities.

The role

As a Working Student in AI/LLM Engineering, you will contribute to the design, development, and assessment of agentic AI systems that convert questions into dependable analytics and automated workflows. The work spans model prompting, tool use, data access, and evaluation, and you will collaborate closely with senior AI specialists on real-world enterprise scenarios.

Key responsibilities

• Design and enhance agent-based workflows for analytics, summarization, visualization, and task automation, including planning, tool/function calls, and self-check mechanisms.
• Create adapters and tools that allow LLMs to interact with internal and external services.
• Support the FastAPI backend by building clean interfaces, adding Pydantic-based validation, and writing tests.
• Develop evaluation methods to track accuracy, latency, and cost.
• Tune prompts, retrieval/context handling, and execution strategies to improve privacy, reliability, and performance.
• Package and ship work in containers using Docker, and collaborate on deployments, Kubernetes, CI, and observability.
• Document technical decisions and share insights with the broader team.

Requirements

• Hands-on exposure to LLMs or AI agents, such as function/tool calling, RAG, MCP, or agent frameworks, along with solid machine learning fundamentals.
• Strong Python ability for production-quality code and data-related work, including typing, testing, and packaging.
• Comfort working with APIs and microservices, ideally with FastAPI, plus Git and containerized development using Docker.
• Good analytical thinking, clear communication, and a proactive approach with a sense of ownership.
• Current enrollment in Computer Science, Data Science, AI, or a similar program, with eligibility to work as a Werkstudent:in in Germany.
• Useful additional exposure includes retrieval/RAG systems, embeddings, chunking, evaluators, Redis, PostgreSQL, Kubernetes, GitLab CI, observability tooling, React/TypeScript, self-hosted or open models, model serving, privacy/GDPR, safety-by-design, and knowledge-graph concepts.

Benefits

• Opportunity to contribute to meaningful, privacy-first AI products used in real enterprise pilots.
• Close mentorship and rapid learning alongside senior AI experts.
• Flexible working hours, remote-friendly setup, and the chance to make an impact from the start.
• Competitive compensation for a Werkstudent role.

Additional information

This position is based in Germany and is structured as a full-time remote working student role. The job emphasizes privacy, enterprise reliability, and deployment flexibility across both on-premise and cloud environments.