AI Engineer at Eclevar Medtech

AI Engineer with a strong focus on applications of LLMs and RAG. Implemented scalable AI solutions using LangChain, LangGraph, Haystack, and vector databases; worked with Llama, Mixtral, GPT across end-to-end pipelines. Conducted data extraction and preprocessing for biomedical articles, implemented retrieval-augmented generation (RAG) techniques to enhance comprehension, and performed model evaluation with faithfulness, precision, and recap measures. Built automation to streamline repetitive tasks; tracked progress with Git and Jira; contributed to task planning and collaboration. Created prompt engineering strategies, integrated APIs and deployed AI pipelines using cloud storage and Vertex AI. Produced interpretable transcripts to support downstream QA and learning tasks.

Internship AI Engineer (NLP/LLMs) at IncoVia GmbH

Internship focusing on NLP/LLMs; explored reproducibility and deep learning methods for NLP/LLMs; evaluated multi-architecture models and contributed to research and development tasks in a corporate environment.

HiWi (Computer Vision) at Friedrich-Schiller-Universität Jena

Student assistant supporting computer vision research; contributed to data collection, annotation, preprocessing, and experiments; assisted in ML model training and evaluation.

Deep Learning Engineer at Max Planck Institute Jena

Developed AI application development for physician consultation assistance; leveraged deep learning models and ML pipelines; contributed to data processing, model training, and deployment in a research environment.

AI Engineer at Elevar Medtech

Remote AI engineering role focusing on applying large language models and retrieval-augmented techniques to healthcare-related AI solutions. Involved in cloud-based deployment, data pipelines, and collaboration with cross-functional teams to deliver scalable AI features.

AI Engineer (NLP/LLMs) Intern at Cowia GmbH

Internship focused on NLP/LLMs at Cowia GmbH in Ilmenau, Germany. Worked on experiments with Llama, Mixtral, GPT, and vector databases to build efficient retrieval systems and explore practical AI applications.

AI Engineer (NLP/LLMs) Intern at Friedrich-Schiller-Universität Jena

Explored reproducibility of deep learning methods using large language models as part of a research project at Jena. Contributed to experiments, data management, and evaluation strategies for LM-based methods.

HIWI (Computer Vision) at Friedrich-Schiller-Universität Jena

Student research assistant role focusing on computer vision tasks, data processing, and supporting DL projects in an academic setting.

Deep Learning Engineer at Max Planck Institute Jena

Deep learning engineering role contributing to research-oriented DL projects at the Max Planck Institute, Jena.

AI Engineer at Eclevar Medtech

Remote AI engineering role focusing on AI-enabled medical technology. Implemented scalable cloud pipelines (Google Cloud Platform) including Cloud Storage, Cloud Functions and Vertex AI for model deployment and data pipelines. Explored retrieval augmented generation (RAG) techniques, evaluated large language models (LLaMA, Mixtral, GPT) for accuracy and faithfulness, and developed end-to-end AI readiness for AI-assisted workflows. Implemented data extraction and preprocessing steps; built an autonomous AI agent for ticket handling via Jira; leveraged FAISS for efficient similarity search and context retrieval; built prompt engineering and improved response quality; integrated with APIs and ensured high faithfulness and precision; used LangChain and GPT-based tools for extensibility.

AI Engineer (NLP/LLMs) Internship at Cowia GmbH

NLP/LLMs-focused internship. Built scalable NLP pipelines, experimented with multiple LLMs for text understanding and generation, and contributed to data pipelines and model evaluation. Implemented tokenization workflows and explored multi-LLM orchestration and evaluation with WER metrics; leveraged FAISS for fast similarity search and context retrieval; contributed to research on model evaluation and deployment.

HiWi (Computer Vision) at Friedrich-Schiller-Universität Jena

Supported computer vision research and AI application development for physician consultation assistance. Contributed to data labeling, preprocessing, experiments, and prototype development; worked with Python, PyTorch, and OpenCV; collaborated in agile team environments and maintained datasets.

Deep Learning Engineer (HiWi) at Max Planck Institute Jena

Deep learning research support, model development and evaluation for medical/vision tasks; built data pipelines and tooling for experiments; contributed to object detection and data processing workflows.

AI Engineer at Elevar Medtech

Developed and deployed AI-based retrieval and LLM-powered assistant features. Leveraged LangChain, Llama, Mixtral, GPT, and vector databases to build scalable AI systems. Implemented data processing and automated workflows to improve productivity.

AI Engineer (NLP/LLMs) Intern at Inovia GmbH

Internship focusing on NLP/LLMs; gained experience with retrieval-augmented generation, fine-tuning, evaluation, and integration into existing ML pipelines.

Deep Learning Engineer (HiWi) at Max Planck Institute Jena

Assisted research in deep learning and computer vision; contributed to AI-related projects, including experiments and implementation for research purposes.

AI Engineer (NLP/LLMs) at PHP Concept

Built a Subsidies AI Service (Python/FastAPI) enabling semantic programme search, subsidy recommendations and application draft/validation workflows. Implemented RAG style retrieval using Qdrant vector search with embeddings/LLMs, including heuristic re-ranking (region, SME, topic) and snippet/context generation for higher-quality results. Designed a structured programme registry with Pydantic, loading YAML-based configurations and optionally enriching records via SQL (Postgres) with environment-based allowlisting. Developed a conversational profile builder to extract company/project details (regex with LLM fallback) to drive more relevant recommendations and reduce manual input. Containerized the stack with Docker Compose and documented local run/testing for quick onboarding and reproducible environments. Created a layered “patch” Docker image that builds on an upstream base image and only overrides the API router which helps in minimal-delta releases and faster, safer CI/CD deplo

AI Engineer at Eclevar Medtech

Built a clinical consultation assistant on GCP (Vertex AI and Cloud Run) that surfaces guideline-backed answers during visits. Pilots cut lookup time from minutes to seconds and reduced post-visit documentation by about 30-35%. Productionized medical ASR by fine-tuning Whisper and Wav2Vec2 on de-identified audio with VAD and domain lexicons. Added LLMs (Gemini, GPT, Claude, Llama) for SOAP-style summaries, medication and allergy extraction and risk flags. Used RAG over a vetted corpus to keep answers grounded with citations. Implemented guardrails including prompt checks, citation requirements, PHI scrubbing and refusal policies. Logged prompts and completions with PII hashing to meet GDPR requirements. Set up evals and observability using RAGAS, task-specific EM/F1 and a lightweight human-rating UI. Prevented regressions and shipped safe A/Bs. Median latency dropped by about 30% using quantized inference with vLLM and response streaming. Moved prototypes to production with containers,

AI Engineer (NLP/LLMs) at Incowia GmbH

Delivered an invoice extraction pipeline that turns scanned PDFs into normalized line-item records using OCR, LayoutLMv3 and an LLM fallback for outliers to lower human review and speed up posting. Fine-tuned BERT-style NER for vendors, addresses, VAT, IBAN and totals. Extracted tables and line items with structure-aware models and confidence gating. Used Mistral as a fallback parser for difficult multi-page invoices, improving recall without a spike in false positives. Cut cost and latency with dynamic batching, mixed precision, safe 4-bit quantization and document-level caching. Maintained stable p95 latency under load and lowered GPU hours. Formalized data and evaluation standards with clear annotation guidelines, inter-annotator agreement checks and CI tests on EM/F1 to block quality regressions. Partnered with product and operations to triage failure modes such as skewed scans, stamp overlays and partial tables. Fed fixes back into training and heuristics for steady quality gains.

Master Thesis - Research in Biomedical NLP (Along Industry Role) at Friedrich-Schiller-Universität Jena

Conducted applied research on Retrieval-Augmented Generation (RAG) for biomedical question answering, working with over 100 scientific papers as source material. Built and optimized dense retrieval pipelines with FAISS and improved precision and recall in complex biomedical text comprehension. Designed evaluation workflows with RAGAS and semantic similarity-based metrics to assess faithfulness and contextual precision of generated answers.