About
Engineer. Author. Occasional observer of systems behaving exactly as designed, just not as expected.
Peter Agambar is a senior software engineer with over four decades of experience building and improving systems across a wide range of industries, including government, healthcare, and AI-driven platforms.
His recent work has focused on cloud-native and data-intensive systems, particularly within AWS environments. This includes large-scale data platforms using Redshift, where he has worked with datasets exceeding 8 billion rows, optimising performance and reducing processing times from minutes to seconds.
His technical background spans .NET, RESTful APIs, microservices, and distributed systems, with a strong focus on performance, scalability, and real-world reliability. He is particularly interested in how systems behave under pressure, where design decisions start to matter, and where complexity can be reduced.
Alongside hands-on engineering, he has led teams, mentored developers, and worked closely with stakeholders to deliver practical, maintainable solutions. His approach is pragmatic—focusing on outcomes rather than theory, and understanding when best practices need to be adapted to fit real-world constraints.
Before his career in software, he served in the Royal Navy and worked across a range of roles, experiences that continue to shape his approach to problem-solving, responsibility, and communication.
He is also the author of the Chapters Are Not Playthings series, where many of the same themes—systems, structure, and unintended consequences—are explored from a slightly more absurd perspective.
Technical focus
Focused on large-scale data platforms, cloud architecture, and system optimisation. Experience includes AWS Redshift, S3, Lambda, microservices, and API-driven systems, with an emphasis on performance, reliability, and maintainability.
Problem: Large-scale analytical system with multi-billion row datasets and poor query performance (3–5 minutes per access).
Approach: Reworked data models, optimised sort/distribution keys, reduced scan overhead, and introduced a hybrid pattern to support both analytical and operational access.
Outcome: Reduced query times to seconds, improved stability, and lowered system load.
Problem: Integrating real-time marine data with predictive analytics and legacy systems.
Approach: Built .NET-based services, integrated IoT and live feeds, and modernised legacy applications to .NET 8.
Outcome: Improved system performance, enabled predictive maintenance and real-time decision support.
Problem: Establishing a new development team and delivering a government payments platform.
Approach: Led team setup, introduced standards, supported Azure B2C integration, and implemented CI/CD pipelines.
Outcome: Delivered a stable platform with improved team capability and consistent delivery practices.
Problem: Fragmented communication systems and need for scalable API integration.
Approach: Designed RESTful microservices using .NET Core, integrated third-party APIs, and improved system interoperability.
Outcome: More reliable communication systems and improved performance across services.