Developed an intelligent chatbot for DeSi restaurant to handle natural dining inquiries with automated menu updates.
Implemented RAG System for handling diverse dining queries
Built pipeline for automatic knowledge base updates with menu changes
Seamlessly integrated with existing restaurant systems
Utilized GitHub Actions for source control and CI/CD
A car sales-focused chatbot powered by LLM that provides answers from FAQs and the cars database
Built from scratch with ADA embedding, Pinecone, and GPT
Fine-tuned pre-trained BERT model for the routing layer
Built text-to-SQL agent from scratch to retrieve data
Utilized Azure DevOps for source control and CI/CD
A GenAI powered tool to assist the SAP team in preparing TS reports, writing ABAP code, and converting ECC to HANA.
Utilized GPT and prompt engineering techniques to prepare TS reports
Automated ABAP code generation based on processed TS reports
Deployed the solution as an API to Azure WebApps
Seamlessly integrated with existing SAP workflows and processes
Served as both technical coach and jury member at an international hackathon with 100+ participants from 24 countries in Freiburg.
Evaluated diverse AI projects ranging from computer vision to LLMs integrated with sensors and robotic systems
One of 7 jury members representing partner organizations, assessing technical innovation and implementation
Collaborated with participants from 24 countries over a 3-day intensive event
Developed an innovative portfolio chatbot integrated with PARSONII, the digital human avatar platform for interactive professional presentation.
Connected custom chatbot with PARSONII avatar system for seamless visual interaction
Built comprehensive response system about internal prodcuts from NTT Data
A face recognition system trained with limited data, enhanced through transfer learning and data augmentation.
Developed baseline recognition system using OpenCV and TensorFlow
Implemented MTCNN for detection and Facebook Deepface/ResNet for recognition
Used transfer learning to improve recognition accuracy with limited data
Combined multiple deep learning approaches for robust performance
Led technical coaching for 50+ participants across 10 teams at the World Conference Center Bonn, focusing on AI implementation and Azure cloud solutions.
Guided teams in implementing AI solutions using Microsoft Azure services
Contributed to event organization and case preparation for participating teams
Provided hands-on support with Azure AI tools and implementation strategies
A GPT-Powered system to calculate carbon footprint and emissions and provide route recommendations.
Utilized OpenAI's GPT and Google Maps APIs to recommend the shortest path and reduce the carbon footprint
Integrated an API directly with SAP to transfer the shipments information
Deployed the model as an API to Azure Web Apps
Feel free to reach out for collaborations, opportunities, or just to say hi!