Zbigniew Mastylo

Project Overview

Led the development of an AI-based search system for an autistic children's school, using advanced embedding techniques to streamline the lesson search process for teachers.

Key Achievements

• Reduced search time by 10x, significantly improving teacher workflow efficiency
• Implemented a user-friendly interface allowing teachers to quickly find relevant lessons
• Utilized SBERT and cross-embeddings for highly accurate semantic search
• Integrated with existing school systems for seamless adoption

Technologies Used

Project Overview

Developed an AI-powered recommendation system that analyzes learner summaries to suggest appropriate lessons for teachers working with autistic children, focusing on specific skill development needs.

Key Achievements

• Created a personalized recommendation engine tailored to each child's unique needs
• Implemented FastAPI backend for high-performance API responses
• Utilized Rivusio for efficient data streaming and processing
• Containerized the solution with Docker for easy deployment and scaling

Technologies Used

Project Overview

Built an AI system that enhances session log notes by automatically generating detailed behavior and lesson narratives based on session data, improving documentation quality and consistency.

Key Achievements

• Automated the generation of comprehensive session notes, saving teachers valuable time
• Integrated behavior data analysis to provide insights on student progress
• Implemented a secure API for seamless integration with existing systems
• Designed a scalable architecture to handle growing data volumes

Technologies Used

Project Overview

Designed and implemented a polymorphic reporting system capable of running both on-demand and scheduled reports with massive scalability and fault tolerance, replacing an outdated Java-based system.

Key Achievements

• Created a pluggable interface making it easy to add new reports to the system
• Reduced report development time from months to days
• Implemented serverless architecture using AWS Lambda and Step Functions
• Built a system with high scalability and fault tolerance capabilities
• Opened new opportunities for users and business teams to request custom reports

Technologies Used

Project Overview

Built fault-tolerant data pipelines and a brand new data warehousing solution for a data startup, enabling large-scale data processing and attracting major clients like Johns Hopkins Hospital.

Key Achievements

• Created a scalable data warehouse that supported ML algorithms for healthcare analytics
• Implemented data pipelines that processed and analyzed hospital, clinic, and doctor performance data
• Helped attract major clients, contributing to the company's 10x acquisition valuation
• Built a system that provided valuable statistics on healthcare service quality

Technologies Used

Project Overview

Designed and implemented a real-time audit platform that tracks user actions, system events, and compliance requirements across a distributed microservice architecture.

Key Achievements

• Reduced compliance reporting time from days to minutes
• Implemented real-time alerting for security and compliance violations
• Built comprehensive dashboards for audit trail visualization
• Designed a scalable architecture handling 10M+ daily events

Technologies Used

Project Overview

Created a standardized API platform that enabled rapid development and deployment of new services, dramatically reducing time-to-market for new features.

Key Achievements

• Reduced new API development time from weeks to days
• Implemented consistent authentication, logging, and error handling
• Created comprehensive documentation and developer tooling
• Designed for horizontal scaling to handle peak loads

Technologies Used

Project Overview

Refactored a critical data infrastructure that was experiencing frequent failures, implementing a robust, fault-tolerant architecture with comprehensive monitoring and recovery mechanisms.

Key Achievements

• Eliminated single points of failure in the data pipeline
• Implemented automatic recovery mechanisms for failed processes
• Added comprehensive monitoring and alerting
• Improved data throughput by 40% while reducing resource usage

Technologies Used

Project Overview

Led a comprehensive transformation of an engineering team promoting well-being, collaborative culture and 80/20 approach. Established culture of trust, quality and 10x mentality.

Key Achievements

• Doubled team velocity within 6 months
• Reduced production incidents by 75%
• Mentored 6 engineers and 2 interns to higher performance levels
• Implemented effective code review and quality assurance processes

Approach

Project Overview

Designed and implemented an automated reporting system that replaced manual processes, saving significant time and resources while improving data accuracy and availability.

Key Achievements

• Saved $1.5MM annually by eliminating manual reporting processes
• Reduced report generation time from days to minutes
• Improved data accuracy and consistency
• Created self-service dashboards for stakeholders

Technologies Used

Project Overview

Developed a specialized search tool for compound curation and discovery, dramatically improving search performance and enabling new research workflows.

Key Achievements

• Reduced search time from minutes to seconds (10x improvement)
• Implemented advanced filtering and faceted search capabilities
• Designed an intuitive user interface for complex search operations
• Scaled to handle millions of compound records

Technologies Used

Project Overview

Researched third-party vendors to enable real-time transactional data replication for a major bank, allowing for processing transactions in active-active mode. Selected and implemented DRNet to provide this critical functionality.

Key Achievements

• Evaluated and recommended DRNet for purchase, which was approved by bank leadership
• Redesigned the existing system to implement active-active transaction processing
• Significantly improved transaction throughput for the bank
• Eliminated balance inconsistencies for business credit card accounts
• Delivered a mission-critical system for a major financial institution

Technologies Used

Project Overview

Architected and implemented a polymorphic system allowing clinicians to test long-term care programs created in a proprietary company pathway language, significantly improving program correctness and validation efficiency.

Key Achievements

• Designed a flexible testing framework adaptable to various program types
• Dramatically reduced the time required to validate each care program
• Improved overall program correctness through comprehensive testing
• Created an intuitive interface for clinicians to use without extensive technical knowledge
• Implemented the solution using modern cloud technologies

Technologies Used

About Rivusio

Cutting-edge Python library for data streaming, transformations, and processing.

GitHub: https://github.com/zbytealchemy/rivusio/

About ZephCast

Lightweight, high-performance messaging library for RabbitMQ, Redis, and Kafka.

GitHub: https://github.com/zbytealchemy/zephcast/