Working Student / Intern – AI Engineering-LLM Systems (Ankara, DACH, remote)

AI Engineering Intern — LLM SystemsGordian Analytics is building AI systems that go beyond simple chatbots and API integrations. We are developing multi-stage LLM pipelines that can understand context, make decisions, generate responses, evaluate their own outputs, and continuously improve based on measured results.We are looking for an AI Engineering Intern who wants to work on real LLM systems and own a meaningful part of the pipeline from design to implementation and iteration.What You’ll Work OnDepending on your strengths and interests, you may take ownership of one of the following areas:Evaluation FrameworkBuild systems that measure whether LLM outputs are actually good. Work with tools such as Playwrite, create evaluation datasets and human-review workflows, and track model performance across different prompts and versions.Intent & Context ExtractionBuild pipelines that extract intent, emotional context, and subtext from text. Use Pydantic and Instructor for structured outputs, test against human-labeled examples, and iterate on reliability.Self-Critique & Quality ControlBuild an LLM-based review layer that evaluates generated outputs before they are used. Develop evaluation criteria, measure how effectively the system identifies poor outputs, and improve the process using tools such as LangFuse.Internal DashboardBuild internal tools that allow the team to monitor pipeline activity and performance. Visualize metrics such as discovery volume, response rates, strategy performance, and model evaluation results using Streamlit, Metabase, React, or Next.js.You will work on a defined component and see your work integrated into the broader system.Tech StackYou don't need to know everything below. Strong Python fundamentals and a willingness to learn are more important.Core: Python 3.11+, FastAPI, PostgreSQL, Redis, Celery / RQLLM: OpenAI / Anthropic APIs, Instructor / Outlines, LiteLLM, Promptfoo, LangFuseData: Jupyter / Marimo, pgvector / QdrantFrontend: Streamlit, React, Next.jsDevOps: Docker, GitHub ActionsWhat We're Looking ForRequired:Comfortable programming in PythonExperience using an LLM API such as OpenAI or AnthropicStrong problem-solving and programming fundamentalsAbility to work independently and communicate clearlyCuriosity about how to evaluate and improve AI systemsNice to have:Pydantic or strongly typed PythonFastAPI or another API frameworkLangChain, LlamaIndex, or similar frameworksPostgreSQL / SQLDocker or CI/CDStreamlit or dashboard developmentPrevious experience building non-trivial LLM applicationsExperience with AI evaluation or experimentationYou do not need years of experience, a perfect GPA, or knowledge of every technology listed above.What You’ll LearnBuilding production-oriented LLM systemsEvaluating non-deterministic AI outputsDesigning multi-stage LLM pipelinesCreating reliable structured outputsUsing observability tools such as LangFuseDesigning experiments and measuring what actually worksWorking with ambiguity and making engineering decisions in an early-stage companyThe First 90 DaysFirst 30 days:Understand the architecture and ship your first component.First 60 days:Take ownership of a defined stage of the pipeline and operate it with human-in-the-loop evaluation.First 90 days:Improve the system based on measured results and begin increasing the level of automation.What We OfferRemote working environmentFlexible part-time internshipApproximately 20+ hours per week3–6 month internshipDirect mentorship and weekly 1:1s with the founderOwnership of a real component of an AI systemOpportunity to work with modern LLM, evaluation, and agent technologiesPotential to continue working with the company as the project develops