Marek – MLOps, Microsoft Azure, Kubernetes

The client was a legaltech company providing advanced research tools to legal professionals. As part of their flagship product, Westlaw, they developed an Intent Resolver system that leveraged natural language processing to interpret user queries more accurately. This improved the precision and relevance of legal search results, enhancing the research experience for attorneys, law firms, and legal departments.

• led the design and implementation of semantic analysis solutions using vector search capabilities;
• developed data collection strategies for the Intent Resolver pipeline using LangChain for orchestration;
• created comprehensive dashboards for product metrics visualization;
• enhanced Python code for AI system event processing;
• implemented vector embeddings for analyzing user queries and improving search accuracy;
• built LangChain pipelines for document processing and semantic retrieval;
• architected end-to-end MLOps pipelines on Azure;
• implemented vector search capabilities using Azure AI Search;
• built RAG systems for improved chatbot responses and document analysis;
• designed microservices architecture for scalability;
• integrated with Azure OpenAI for advanced NLP capabilities;
• engineered solution for analyzing SEC EDGAR filings using LLMs;
• created entity extraction systems for financial data;
• implemented user feedback loops for continuous improvement;
• performed data mining using Snowflake Cortex AI;
• integrated multiple data sources to create unified analytical views.

The client was a technology consultancy delivering custom software and IT services to enterprise clients across Europe. They specialized in supporting digital transformation efforts by providing multidisciplinary engineering teams, integrating tailored software solutions, and enhancing business operations across sectors such as finance, healthcare, and automotive. The project itself was an enterprise MLOps platform

• architected and implemented end-to-end MLOps pipelines on Azure;
• optimized data processing workflows using Databricks;
• implemented CI/CD for ML model deployment;
• designed data storage solutions with Parquet for efficient processing;
• automated model retraining and versioning;
• reduced cloud infrastructure costs by optimizing resource allocation.

The client was a conversational AI company specializing in chatbot and voicebot solutions for businesses. Leveraging expertise from Stanusch Technologies, they developed an omnichannel platform tailored for Polish and Slavic languages, enabling organizations to automate customer interactions, enhance internal processes, and drive digital transformation across various industries.

• designed and implemented complex, scalable chatbot conversation systems;
• optimized knowledge base storage for complex NLP queries;
• led Deep Learning projects for speech recognition (ASR) and sentiment analysis;
• implemented Hidden Markov model and Trie algorithms for fast NLP search;
• preprocessed customer knowledge bases using LLMs;
• built applications handling concurrent conversations and large knowledge bases;
• analyzed time-series data for ASR and chatbot accuracy monitoring.

The client was a technology consultancy specializing in custom software development and AI-driven automation for enterprise clients. They provided end-to-end services including team augmentation, system integration, and intelligent process automation. Their solutions supported businesses in sectors such as e-commerce, legaltech, and manufacturing, helping them scale operations, reduce costs, and accelerate digital transformation.

• designed system architecture for an AI-powered monitoring solution;
• enhanced smart monitoring using Generative AI and LLM-based systems;
• integrated NLP for knowledge management;
• designed MLOps data pipelines for incident detection and analysis;
• integrated OpenAI and implemented RAG vector storage;
• developed sophisticated LangChain pipelines for orchestrating complex AI workflows;
• led Deep Learning projects, developing ML models for automatic incident detection;
• created an Autonomous Support Agent for system maintenance using LangChain and LangGraph;
• leveraged LLMs for incident classification and root cause analysis;
• built custom LangChain agents for automated troubleshooting and system optimization.

The client was a global enterprise software company specializing in integrated business applications for finance, supply chain, human resources, and customer experience. They developed a cloud-based ERP platform enhanced with AI-driven tools, enabling organizations to streamline operations, make data-informed decisions, and adapt to evolving market demands. Their solutions served businesses of all sizes across various industries, supporting digital transformation and operational efficiency.

• developed ML-based system for comprehensive customer profiles;
• created deep neural network models for product recommendations;
• built a GDPR-compliant data processing system;
• designed a unified data platform across multiple sales channels;
• led AI capabilities integration with SAP's intelligence suite;
• analyzed large datasets of customer behavior patterns;
• implemented scalable models for customer profiling.