Hi, I'm Tymon 🚀

Built end-to-end features for a document-processing platform across frontend and backend. Created a contextual data provider driven by extraction outputs from mortgage documents, extending product capabilities and enabling smoother user integrations. Optimized CI/CD and deployment (containerization, release automation), shortening release cycles and improving reliability and operational efficiency.

Built an end-to-end document LLM extraction pipeline combining layout-aware OCR (Tesseract/EasyOCR, PaddleOCR, PyMuPDF) with LLMs (Qwen, Deepseek, GPT-OSS family) to collect notarial entities into schema-constrained JSON. Optimized performance by migrating from Ollama to production LLM backends like vLLM and SGLang, architecting a message-driven OCR-to-LLM workflow with RabbitMQ and a Python worker bridge, and containerizing services with Docker while tuning runtime parameters to balance quality and latency on DataCrunch infrastructure.

At Point72, I focused on developing and optimizing pipelines, building and maintaining APIs, and utilizing AWS services to create scalable and efficient solutions. I engineered data-driven dashboards using Dash, integrating them with backend services and APIs to deliver actionable insights. My work involved designing and implementing robust data pipelines, ensuring seamless data flow and processing across various platforms, while leveraging AWS to enhance system performance and reliability. I also played a key role in API development, enabling smooth integration between services and enhancing overall system functionality.

I engineered and implemented a Python-based NLP system for mKanon using FastAPI, Spacy, Morfeusz2, and scikit-learn for data classification and dimension reduction, while also developing a task analysis dashboard to optimize internal processes through clustering techniques. My work on the "CCL Component Library" involved adding features to key components, building large components, and integrating them with business logic. Additionally, I developed SPAs connected to API endpoints for mBank clients, resolved frontend security vulnerabilities, and introduced new DOM handling logic for mKanon. I automated visual testing for Design System components and integrated automation tools like Gmail API and GCP. Furthermore, I led technical interviews, managed project planning, and drove process improvements that enhanced team collaboration and overall efficiency.

At Persolozized Solutions, I specialized in commerce development, front-end engineering, and the application of AI solutions across various sectors. My work involved creating responsive and scalable front-end interfaces, optimizing user experiences, and integrating advanced AI-driven features to enhance functionality and decision-making processes in commercial platforms. I played a key role in developing customized solutions tailored to diverse business needs, leveraging AI technologies to drive innovation and streamline operations. This included building efficient e-commerce systems and implementing AI models that improved business outcomes across different industries.

Developed commercial web applications for the banking sector using Angular 11, JavaScript, and TypeScript. Enhanced office organization during the pandemic by optimizing an internal web application.