Automated content generation system ("NISI"): developed an AI-powered system to automate the customer journey, sales and corporate branding. Created AI-generated articles from customer interviews using LangChain, LangGraph, OpenAI API, which led to 54 new user sign-ups and facilitated customer feedback and product discovery. Technologies used: Python, Git, CI/CD, Docker, Machine Learning, LLM, LangChain, LangGraph, LangSmith, OpenAI, OpenSource Development, Google Cloud, ChatGPT.
Business value estimation tool: developed a tool to analyze repositories and assign business values, supporting prioritization and risk management. The tool consists of different AI tools combined in a graph. Technologies used: Python, Git, CI/CD, RESTful API, Docker, Machine Learning, LLM, LangChain, LangGraph, LangSmith, OpenAI, Cybersecurity, OpenSource Development, Google Cloud, ChatGPT.
Automated risk mitigation: implemented automated risk mitigation using AI tools, eliminating several risks in the company's product. Technologies used: Python, Git, Docker, CI/CD, Machine Learning, LLM, LangChain, LangGraph, LangSmith, OpenAI, Neo4j, Cybersecurity, OpenSource Development, ChatGPT.
RAG, Graph-to-Chat/Chat-to-Graph system: designed and implemented a system that enables bidirectional mapping between chat dialogues and the knowledge graph as RAG. This project can extract and integrate conversational data in both directions into a dynamic graph structure, providing an efficient user experience for managing cybersecurity data and dialogues. Technologies used: Python, Git, Docker, Machine Learning, LLM, LangChain, OpenAI, Numpy, Cybersecurity, Neo4j, OpenSource Development, Google Cloud, ChatGPT.
Optimized AI workflows: integrated AI tools and optimized cost efficiency, deploying solutions on Google Cloud Run.
Code review and quality assurance: reviewed pull requests and refactored code to improve quality and maintainability.
Team training and leadership: conducted training sessions on AI tools and workflows, promoting continuous learning.
Nov 2022 - Jul 2023
9 months
Machine Learning Researcher/Developer
Siemens Energy
Developed an ML model with PyTorch and PyTorch Geometric to automate high-voltage configuration processes, reducing configuration time by up to 30 hours per instance.
Achieved 96% accuracy in link prediction for domain-specific knowledge graphs.
Collaborated with cross-functional teams for data collection and domain understanding.
Migration tool from Polarion to Azure DevOps: developed a migration tool to transfer data of epics, stories, sprints, etc. from Polarion to Azure DevOps using Python and Azure REST API. Technologies used: Python, RESTful API.
Automated testing systems: developed advanced automated test systems for functional and smoke tests to increase software quality. Technologies used: Git, Python, CI/CD, C#.
Monitoring system: developed a Python-based monitoring system for Windows servers, Windows services, and other interfaces (between applications). Technologies used: Python, RESTful API, CI/CD.
Software adapter: designed and implemented a software adapter in C# to extend system functionality. Technologies used: C#.
Automated tasks: implemented a framework for automated tasks, user management, and process monitoring. Technologies used: C#.
DevOps
Scalable Enterprise Web Service Development – Azure & Terraform.
Designed and developed a highly scalable and modular web service with Azure and Terraform, enabling seamless expansion and enterprise-wide deployment.
Infrastructure and system architecture design, ensuring a secure, reliable, and efficient structure that can be sold as a multi-tenant enterprise solution.
Built a full booking system, including frontend, backend, API integrations, database management, and secure server deployment.
Implemented automated infrastructure deployment with Terraform, optimizing deployment, scalability, and maintainability.
Ensured high security standards, including authentication, encryption, and best practices compliance for cloud-based applications.