[ 01 ]
TourFlow AI
Deterministic State Machines for Non-Deterministic Models
Python strategy patterns for Crown Uncovered, a London AI tour.
GitHub// SYSTEMS & AI ARCHITECTURE
Franco
Begue_
Systems & AI Architecture Lead Architect at Personatio
I build deterministic state machines for stochastic AI models, high-throughput Rust data pipelines, and distributed Kubernetes topologies. Outside of my work at Personatio, I write in-depth technical teardowns and maintain open-source systems.
01 / SELECTED SYSTEMS
The Toybox
02 / FIELD NOTES
Writing
2026.08.19
Why LLM Mega-Prompts Fail at Scale: Building Deterministic State Machines with the Strategy Pattern
A TourFlow teardown: make language models expressive at the edge, while an explicit state machine owns the system's truth.
2026.08.12
Streaming PostgreSQL to Parquet at Wire Speed: Zero-Copy Serialization in Rust
How postgres_to_parquet avoids row-shaped bottlenecks by treating extraction as a streaming binary pipeline.
2026.08.05
Kubernetes Edge Cases: Designing Failover Boundaries for High-Concurrency Microservices
Resilience is not a retry policy: it is an intentional set of boundaries around failure domains, queues, and state.
03 / SPECIAL CASES
// UNUSUAL
SYSTEMS
Most of my focus is on building core systems at Personatio. Occasionally, another problem earns a place at the workbench.
Architecture work where the stakes are real, the constraints are interesting, and the answer has to hold.
Inquiries: frank@personatio.com or LinkedIn