Software Engineer

12/2023 – offen

• Developed a project to predict HS6 classification based on product description:
• Initially developed multi-label TextClassifier model to classify invoice line fields
• After gpt-4o appeared used OpenAI API to create product catalogs.
• Used HuggingFace, Pytorch, scikit-learn libraries and openNMT and BERT models.
• Used AWS Sagemaker, Bedrock, OpenAI and langchain

• Developed model to predict invoice payment time
• Used Xgboost, scikit-learn, pandas, numpy and MLXtend libraries
• Used OpenAI to generate product catalogs based on invoices lines.
• Used langchain and OpenAI libraries

• Developed over 700+ models to predict customer specific invoice codes.
• Experimented with Fasttext, DecisionTrees, LogisticRegression, NaiveBayes, RandomForest, Xgboost and
LightGBM.
• Deployed DecisionTrees models for cost-effective hosting.
• Used Natural language text processing (CountVectorizer, Tfidf Vectorizer and embeddings) for feature
engineering from text information.
• Automated model training through sagemaker jobs

• Worked on model to predict mappings between external invoice lines field names and internal mappings
• Experimented with FastText and BERT
• Developed Multi-label (not to be confused with multi-class) models.
• Used Pytorch, Huggingface, skmultilearn, scikit-learn, pandas, numpy

Multi-Agent System for Business Firewall Rule Management
• Project Overview:
• Designed and implemented a multi-agent system to streamline the migration of legacy firewall rule
code.
• Key Components:
• Planner Agent: Orchestrated the overall migration workflow by assigning tasks and coordinating
between agents.
• Groovy Agent: Interacted with legacy Groovy code to extract and interpret existing firewall rules.
• JSON Agent: Assisted developers by generating and validating new JSON code, ensuring a smooth
transition from legacy systems.
• Technologies Used: Azure OpenAI, Chainlit, LangChain, and RAG.

Amazon Web Services (AWS), Docker, Generative KI, Git, GPT, Gradient Boosting, Kubernetes, Large Language Models, Pandas, Python-Programmierer, Pytorch, Scikit-learn

10/2022 – 2/2024

• Worked on a LLM project to generate Jira stories based on project documentation.
• Used the following LLM models: AzureOpenAI and AzureOpenAIEmbeddings
• Used the following libraries: Langhcain, chromadb

• Worked on a recommender system
• Used multiple models: LightFM, LightGBM, and Ensemble Models of these 2, GMF, MLP, NMF.
• Did Feature Engineering, negative sampling, Hyperparameter optimization on models
• Data Visualization
• Drift Analysis

• Designed ETL pipelines using Azure Data Factory and Databricks.
• Developed PySpark scripts and optimized Spark jobs on Databricks.
• Ensured data security with RBAC and encryption
• Resolved schema, deduplication, and performance challenges.

• Worked on Presales Proposals

Data Science, Databricks, Flash-Programmierer, Jira, Microsoft Azure, Pandas, Scikit-learn, SQL, Tensorflow

3/2021 – 10/2022

Technologies used: Python, Scikit-Learn, XGBoost, Pyspark, Pandas, Databricks, Numpy, Shap, Matplotlib, Docker, Tableau, Azure DevOps
Developing price sensitivity machine learning model.
Feature Engineering
Picking appropriate data representation methods.
Identifying differences in data distribution that affects model performance.
Collaborating with Data Engineering Team in order to get the data needed for the model.
Calculating feature importance improve model explainability.
Creating dashboards in Tableau for data visualization.
Model training, testing, debugging and deploymentTechnologies used: Python, Scikit-Learn, XGBoost, Pyspark, Pandas, Databricks, Numpy, Shap, Matplotlib, Docker, Tableau, Azure DevOps Developing price sensitivity machine learning model. Feature Engineering Picking appropriate data representation methods. Identifying differences in data distribution that affects model performance. Collaborating with Data Engineering Team in order to get the data needed for the model. Calculating feature importance improve model explainability. Creating dashboards in Tableau for data visualization. Model training, testing, debugging and deployment
Python (Programming Language), Data Science an

Apache Spark, Data Science, Databricks, Docker, Flash-Programmierer, Machine Learning Engineer, Microsoft Azure, MLOps, Pandas, Python, Scikit-learn