Maciej – Python, LLM, OpenAI

Development and enhancement of an AI-powered chatbot tool designed to assist customer service agents in providing efficient and accurate support for diagnostic instruments. The tool helps automate routine inquiries, offers real-time recommendations, and improves response quality, ultimately supporting smoother troubleshooting and better customer experience within the healthcare diagnostics domain.

• Refactored POC to Clean Architecture, decoupled 3rd-party dependencies, implemented test suite and quality checks;
• Built AI copilot with RAG for 500+ users on Azure;
• OpenAI/Langchain/Langraph for Roche diagnostic troubleshooting;
• Developed intelligent agents with tool-calling for multi-step support workflows;
• Led monolith-to-microservices refactoring with frontend/backend separation;
• Improved copilot accuracy through prompt engineering workshops with domain experts;
• Increased test coverage 40% → 80% with TDD and CI/CD quality gates;
• Conducted 6+ AI workshops on Cursor/Copilot, boosting team productivity;
• Mentored junior engineers on clean architecture and AI best practices.

A centralized REST API solution that connects an Airtable database with major advertising platforms (Meta, Google, TikTok) to enable automated creation and management of ad campaigns. The project aims to eliminate repetitive manual tasks for media buying teams, ensuring faster campaign launches, improved data consistency, and better scalability across multiple marketing channels.

• Solo-architected automation platform integrating Airtable, Meta, Google Ads, and TikTok APIs, reducing campaign creation from hours to instant;
• Built an extensible Django REST API with a provider pattern for easy platform additions and event-driven Celery tasks;
• Implemented a comprehensive test suite with 90% coverage, including unit and integration tests;
• Deployed on AWS with enterprise security, rate limiting, and API documentation;
• Created extensive technical documentation and conducted knowledge transfer sessions with the company's full-time developer for a smooth handoff.

A specialized software solution designed to manage, maintain, and optimize Change of Line (CoL) processes within pharmaceutical manufacturing laboratories. The project aims to streamline production line adjustments, ensure compliance with industry standards, and enhance overall operational efficiency by automating routine tasks and providing robust monitoring and reporting capabilities.

• Implemented async Django features, reducing API response times;
• Optimized critical API endpoint, cutting latency across the enterprise application;
• Reduced CI/CD build times through Docker optimization;
• Introduced type hints and async patterns to a large Python codebase.

A large-scale platform providing global edge cloud services, a high-performance content delivery network (CDN), and secure hosting solutions. The project focuses on delivering low-latency content distribution, reliable cloud infrastructure, and scalable hosting to support businesses in optimizing their online presence and performance worldwide.

• Built a unified logging package deployed across 6 microservices (2 Django, 4 FastAPI);
• Developed statistics API processing traffic data for real-time dashboards;
• Integrated ML content moderation models via REST API;
• Set up Grafana monitoring with health and performance dashboards;
• Introduced OpenAI and CodeRabbit for automated code reviews.