10.03.2025 aktualisiert
verifiziert
Premiumkunde
100 % verfügbarKI ML LLM Entwickler Python GPT RAG Data Science Computer Vision OpenAI Deep Learning
Berlin, Deutschland
Weltweit
Mathematik B.Sc. + M.Sc. (1.0), Physik B.Sc. an HU BerlinSkills
Computer Visiondeep leariningImage ProcessingNeural NetworkOCREmbeddedPythonC++MatlabOpenCVNumPyTensorflow PyTorchDocker PandasSQLawsLinuxCUDAJetson NanoArduinoRaspberry PiARM CortexRTOSAT - megaPytestPoetryFlake8MyPyPylintATMEL StudioNotfallmanagementLarge Language ModelsGPTChatGPT4LLAMAOpenAILangchainPrompt EngineeringHugging FaceEmbedding ModelsKILThelmThruthfulQAAzureAmazon AWSBERTLamdaTransformerRoBERTaspacyEspnetKALDISpeechbrainmlflow
Specialisations
Machine Learning, Cloud Computing, NLP, OpenAI, GPT4, OCR,
Neural Networks, Computer Vision, Image Processing, Embedded,
Time Series Analysis
Programming
Python > 11 years
C++ & C > 6 years
Matlab > 8 years
Azure > 3 years
AWS > 4 years
Software & Tools
Scikit-learn, OpenCV, Numpy, Tensorflow, PyTorch, PyTesseract, Docker, Pandas, SQL, Spark, AWS Sagemaker, Linux, CUDA, C++, Git, Power BI
IoT & Embedded
Jetson Nano, ARM Cortex, AT- mega
Production Code
Unittest, Pytest, Poetry
Machine Learning, Cloud Computing, NLP, OpenAI, GPT4, OCR,
Neural Networks, Computer Vision, Image Processing, Embedded,
Time Series Analysis
Programming
Python > 11 years
C++ & C > 6 years
Matlab > 8 years
Azure > 3 years
AWS > 4 years
Software & Tools
Scikit-learn, OpenCV, Numpy, Tensorflow, PyTorch, PyTesseract, Docker, Pandas, SQL, Spark, AWS Sagemaker, Linux, CUDA, C++, Git, Power BI
IoT & Embedded
Jetson Nano, ARM Cortex, AT- mega
Production Code
Unittest, Pytest, Poetry
Sprachen
DeutschMutterspracheEnglischverhandlungssicher
Projekthistorie
Project: LLM based Journalistic Agent
- Develop personalized content recommendation tool for media articles via LLMs, Embedding models, RAG.
- Develop self-prompting agent-mechanism pipeline via Langchain Framework.
- Develop Metrics for LLM comparison and hallucination detection, getting LLMs with less hallucination.
- Integrated LLM in a streamlit-playground by integrating LLMs in an AWS Docker Endpoint accessing via FastAPI.
Project: Patient Data interaction with LLM
Development of an LLM language assistant on the iPad for daily medical practice.
• The language assistant summarily stores the spoken information in the patient's file.
• Development of OCR on iPad for converting a photo into text, storing it in the patient's file.
• Interaction with patient documents via LLMs and spoken language.
• Developed LLM finetuning pipeline via Langchain Framework.
• Fine-tuned non-proprietary LLM models for specific use cases
• Developed solutions via attention based metrics how much the LLM is hallucinating, getting LLMs with much less hallucination
• Integrated LLM use cases in a playground of the customer by integrating LLMs in a AWS Docker Endpoint and giving access via REST API
for the connection with the Frontend
Development of an LLM language assistant on the iPad for daily medical practice.
• The language assistant summarily stores the spoken information in the patient's file.
• Development of OCR on iPad for converting a photo into text, storing it in the patient's file.
• Interaction with patient documents via LLMs and spoken language.
• Developed LLM finetuning pipeline via Langchain Framework.
• Fine-tuned non-proprietary LLM models for specific use cases
• Developed solutions via attention based metrics how much the LLM is hallucinating, getting LLMs with much less hallucination
• Integrated LLM use cases in a playground of the customer by integrating LLMs in a AWS Docker Endpoint and giving access via REST API
for the connection with the Frontend
Project: Deploy and optimize OpenAI services
Deploy and optimize OpenAI services for a big customer. Cloud migration in Azure and Azure Cognitive Services. Deploy GPT4, advanced prompt engineering, GPT4 fast API computation with vectorized embeddings using ML. Project ongoing.
Tech Stack: NLP, Azure Cloud, OpenAI, GPT4.0, Python, Prompt Engineering, Gitlab CI/CD, Azure DevOps.
Deploy and optimize OpenAI services for a big customer. Cloud migration in Azure and Azure Cognitive Services. Deploy GPT4, advanced prompt engineering, GPT4 fast API computation with vectorized embeddings using ML. Project ongoing.
Tech Stack: NLP, Azure Cloud, OpenAI, GPT4.0, Python, Prompt Engineering, Gitlab CI/CD, Azure DevOps.
Portfolio
Depth Recognition
Tiefenerkennung via KI und Waveshare Dualcam.
https://www.dropbox.com/s/7qj4sgfh4bu27xz/smart_aal_portfolio.pdf?dl=0
Face Recognition
Gesichtserkennung via Deep Learning
https://www.dropbox.com/s/7qj4sgfh4bu27xz/smart_aal_portfolio.pdf?dl=0
Fall Detection
Fallerkennung via Human Skeleton detection und Deep Learning.
https://www.dropbox.com/s/7qj4sgfh4bu27xz/smart_aal_portfolio.pdf?dl=0
Human Identification
Personen-tracking, Personen-Wiedererkennung, Security
Fall Detection
Fall Detection
Event Annotator
Tool for video event annotation, e.g. can annotate the Fall-events or Water drinking events.
In Out detection
Detects if a person is going out.
Bacteria Detection
Detection of bacteria on microscopic images, 99% accuracy of detection.
Cancer Detection MRI
Development of a neural network-based algorithm to detect, classify and segment breast tumors in mammography X-ray images to improve radiologists' performance in breast cancer screening.
Image Aesthetics
Deep Learning model to predict the aesthetic quality of images, used explicable AI and genetic algorithms to achieve high accuracy.
Robot Arm
Prediction of a torque very accurately. Determination of the optimal trajectory for the robot arm. I have developed a custom neural network to solve this optimization problem using physically modeled helper computations.
Text Evaluator LLM
Bullet Point Evaluator App via LLM RAG Embedding Models Langchain