Mikołaj Zapalski

• Designed and optimized LLM infrastructure to support core features like token management, retries, stream parsing, agent failure handling, and snapshot deployments. Integrated new features into the AI backend with a focus on reliability, scalability, type safety, and clean, maintainable code, driving significant product growth. • Authored research reports to guide data-driven decisions, implementing paragraph-level, hallucination-free citations, infinite-context document representation, advanced embeddings, output validation, and RAG pipelines tailored to legal datasets. Regularly applied SOTA research to enhance product capabilities, rigorously tested through a custom evaluation framework to ensure performance, reliability, and scalability. • Reduced the error rate from 2.5% to 0.02% (100x improvement), significantly improving reliability and user satisfaction. Collaborated with stakeholders in high-pressure 'war room' settings, iterating rapidly on user feedback to secure €500k+ deals.

• Integrated and maintained machine learning models for a holistic wellness mobile app, including a personal LLM assistant, in-app content recommendation system, and semantic search engine. • Designed and deployed efficient ML data pipelines for user activity scoring and content retrieval modules, optimizing app performance. • Leveraged Python expertise (pandas, NumPy, PyTorch, FastAPI) to deploy machine learning models on AWS infrastructure, integrating OpenAI API and wearables API with MongoDB for advanced functionalities and user data management.

• Designed, delivered, and owned numerous internal tech solutions (web scrapers, data pipelines, automation tools) that streamlined workflows and boosted efficiency. • Developed a global contractor payroll system, automating monthly processing for 400 employees ($50k/month), including transfers, invoicing, and P&L reporting. • Successfully implemented a 70% efficiency improvement through automation of contractors' legal paperwork pipeline. • Proposed the exploration of OpenAI API for data augmentation, investigating its potential to increase the size of the client's CRM database by 15%. • Utilized Python, Pandas, Jupyter Notebook, JavaScript, Zapier, and Google Sheets to deliver effective solutions.

• Managed hundreds of multimodal datasets (vision, tabular, text, audio), ensuring data quality through robust annotation, cleaning, and validation processes. • Developed and implemented data quality checks to improve data integrity for machine learning projects, while iteratively optimizing the data annotation process, resulting in a 10x improvement in project delivery time. • Transitioned to a combined role encompassing project and operations management, ensuring smooth data pipeline execution and client satisfaction. • Led the annotation team, gathered and implemented client requirements and feedback, resulting in a significant improvement in model performance and accuracy.

• Provided technical support and incident management for IT systems and peripherals, effectively resolving issues and monitoring & responding to alerts. • Prioritized tasks to minimize global infrastructure disruptions and enhance user experience.

• Seasonal activity in the garden and home automation sectors.