Automated Stock/Crypto Trading System

Automated trading system using Playwright for data extraction, web search with cron jobs, Claude/Codex for analysis, and sentiment/fear analysis for decision making.

Skills: Playwright, Web Scraping, Cron Jobs, Claude API, Codex, Sentiment Analysis, Fear & Greed Index, Trading Automation.

Contextual AI Coach

Browser extension to make daily work easier for colleagues by using the 'key/experienced user' knowledgebases.

Skills: N8N, Playwright, MCP Servers, RAG, Browser Extension.

Low Code Workflows with Power Platform

Creating an agentic application via Power Platform (Copilot Studio and Power Platform).

Skills: Power Platform, Copilot Studio, Low Code Development.

Procurement Agent

Creating an agentic RAG chatbot using Django framework to extract meeting information and audio files and generate arguments for procurement negotiation with different user history and web search. Using vector DB (Milvus, Qdrant) with hybrid search.

Skills: RAG, Agentic Workflow, Vector DB, Hybrid Search, SEARXNG, Django.

Fine Tuning Text to Image Models

Fine-tuned a Text to Image model for internal usage.

Skills: Fine Tuning, Large Language Models (LLM), Machine Learning, Python, TensorFlow/PyTorch.

LLM Fine Tuning

Fine-tuned a Large Language Model (LLM) for specific tasks to improve performance and accuracy.

Skills: Fine Tuning, Large Language Models (LLM), Machine Learning, Python, TensorFlow/PyTorch.

ASR Fine Tuning

Fine-tuned an Automatic Speech Recognition (ASR) model to enhance transcription accuracy.

Skills: Fine Tuning, Automatic Speech Recognition (ASR), Machine Learning, Python, TensorFlow/PyTorch.

Live Subtitling

Used Whisper ASR to create English translations for German videos at 5-second intervals.

Skills: Whisper ASR, Natural Language Processing (NLP), HTML, Large Language Models (LLM), VAD (Voice Activity Detection).

Multi-Agent Chatbot

Developed a chatbot using multi-agent systems integrating real-time data and routing.

Skills: Large Language Models (LLM), LangChain, Langsmith, Google Cloud Platform (GCP), Vector Databases, Indexing, Multi-agent Systems.

Advertisement Generator

Created a tool for generating short advertisement videos.

Key Features:
- Text parsing from URLs
- User interaction and voiceovers
- Image and clip matching with keywords using GPT-4 Vision and Turbo
- Image analysis and voice cloning

Skills: Computer Vision, Vision Embedding, GPT-4, Custom GPTs, Voice Cloning, Multi-Agent, LlamaIndex, LangChain, Langsmith.

Protagonist Detection in Media

Developed a system to identify the main character in videos and detect close-up shots and aesthetic scenes for thumbnails.

Skills: Computer Vision, Large Language Models (LLM), Face Detection, Actor Recognition, Face Clustering, Dimension Reduction, Aesthetic Ranking, Shot Type Detection, Google Cloud Platform (GCP).

Silence Detection

Implemented silence detection to enhance shot/scene detection in videos.

Skills: FFmpeg, Large Language Models (LLM), Whisper ASR, Google Cloud Platform (GCP), Microsoft Azure.

ChatBot with RAG

Key features include internal data chat, multi-format data uploads, web search and summary, user feedback, and multi-agent function calling.

Skills: LlamaIndex, RAG, Mistral7B, OpenSource LLM, Streamlit, Generative AI, OpenAI API, Microsoft Azure, Google Cloud Platform (GCP), Document Chunking, K-Nearest Neighbors (KNN).

Black Frame Detection

Detected black frames in videos and provided timestamps and black frame levels.

Skills: FFmpeg, Google Cloud Platform (GCP), Computer Vision.

Aesthetic Ranking

Evaluated images for aesthetic quality using a CNN-based model and clustered similar frames.

Skills: Convolutional Neural Networks (CNN), Dimension Reduction, UMAP, HDBSCAN, Google Cloud Platform (GCP), K-Means Clustering, Computer Vision, Vision Embedding.

Face Detection

Detected faces in video frames and identified emotions, gender, and age for thumbnail classification.

Skills: Computer Vision, Convolutional Neural Networks (CNN), Google Cloud Platform (GCP).

Topic Modelling

Clustered audio/text/video datasets by topic in multiple languages.

Skills: BERT, Bertopic, Natural Language Processing (NLP), Hugging Face Hub, Sentence Embedding, Language Detection, Large Language Models (LLM), Python, KeyBERT, K-Means Clustering, Dimension Reduction.

KI Interview

Developed an AI-driven interview system showcased on RTL.

Skills: Text-to-Speech, ChatGPT, Generative AI, Speech-to-Text, Microsoft Azure, Microsoft Azure Machine Learning, Azure Functions.

Adult Content Classification

Classified adult content in text/audio/video datasets using OCR.

Skills: Optical Character Recognition (OCR), Natural Language Processing (NLP), Hugging Face Hub, Language Detection, SpaCy, Sentence Similarity, Text Embedding, Regex, Whisper ASR.

Kids Content Classification

Classified kids/juvenile content from text/audio/video datasets.

Skills: Optical Character Recognition (OCR), Automatic Speech Recognition, Text Embedding, Hugging Face Hub, SQL, Regex.

Festivity Detection

Detected festivities in Germany using audio and text datasets.

Skills: Regular Expressions, Sentence Embedding, Text Embedding, Hugging Face Hub, Sentence Similarity, Language Detection, SpaCy.

Language Detection

Identify languages in datasets for recommendation systems.

Skills: Natural Language Processing (NLP), Tokenization, Word Embeddings, Text Embedding, Hugging Face Hub, Poetry.

Keyword Key-phrase Extraction

Extract keywords and phrases from text datasets.

Skills: KeyBERT, Natural Language Processing (NLP), Tokenization, SQL, Language Detection, Text Embedding, SpaCy.

Master Students Project Teaching Assistant

Guide master students on NLP projects like Hate Speech and Sentiment Analysis.

Skills: Analytical Skills, Automatic Speech Recognition, Leadership, Team Leadership, Management, Educational Leadership, Organizational Leadership, Team Building, Machine Learning Algorithms, Supervised Learning, Applied Machine Learning, Human-in-the-Loop Machine Learning, Neuro-Linguistic Programming (NLP).

Churn Prediction (Master Thesis)

Predict churn for a manufacturing company using big data.

Skills: Analytical Skills, Azure Functions, Data Analysis, Data Science, Dimension Reduction, Flask, Git, Machine Learning, NLTK, Churn Prediction, Cluster Analysis.