Systems I've built when the hard part is the backend. Ones you can run so reliably that ops feels boring.
Evidence your team can replay. A Rust harness for realistic system and integration tests against real resources (databases, event buses, etc), so teams get replayable evidence that enables them to trust in a release version before production. Take one or two containerized microservices and stress them with your custom scenarios in reproducible ways.
Rust, Tokio, testcontainers, Kafka, Postgres, Redis
Foundation for a multi-service Rust backend: BFF and domain services over gRPC, Kafka events with a transactional outbox, and per-service Postgres. Built so investing signal detection can scale without a big-ball-of-mud monolith.
Rust, gRPC, Kafka, PostgreSQL, Redis, Docker
A Smalltalk-like DSL that compiles to async Rust actors. Write classes and methods in .s sources; the compiler generates state types, message enums, and actor run loops targeting Tokio, Monoio, or Glommio.
Rust, DSL, Tokio, Monoio, Glommio, actors
lightsnap
Active development
Compressed state snapshots for Solana ZK Compression tests: capture ZK-compressed state from light-program-test, render readable tables, serialize to JSON, and diff before/after runs without hand-reading Merkle trees.
Rust, Solana, Light Protocol, ZK compression
Schema-less persistence for Pharo/Squeak: models you can save without OR impedance (no instVars ceremony, no getters and setters required). Still community-maintained.
Smalltalk, Pharo, persistence
Full-stack web framework for Pharo aimed at consultants and startups who need a modern HTML5 demo path with built-in persistence and productivity tooling.
Smalltalk, Pharo, full-stack