I design and ship deterministic, supply-chain-safe systems in Rust — focusing on performance, correctness, and architectures that eliminate failure modes rather than mitigate them.
4+ years building and deploying Rust across distributed systems, storage engines, embedded targets, and cryptographic primitives. I care deeply about mechanical sympathy, auditability, and long-term maintainability.
cargo-rail
Monorepo orchestration for serious Rust workspaces.
- Replaces
cargo-hakari,cargo-udeps,cargo-machete,cargo-release,git-cliff,Google Copybara,andcargo-msrv - 190+ stars · 14 total dependencies
- Used and tested on
tikv,tokio,polars,helix,meilisearch,ripgrep - Official GitHub Action
rscrypto
Pure-Rust cryptographic primitives with zero C dependencies.
- High-performance CRC implementations (16/24/32/32C/64XZ/64NVME)
- Pure Rust Blake3 impl across all supported platforms at parity with or surpassing the official Blake3 implementation.
- Portable across x86-64, ARM64, RISC-V, s390x, POWER9/P10
- Accepted by IBM for CI/CD usage on mainframe architectures
These projects exist because I needed them for my own work.
I’m building a distributed storage engine at @alias-research, approaching its first public release in the next few months.
The system is centered around a novel epoch-based reclamation primitive that enables:
- Sub-10ns MVCC operations
- Deterministic O(1) garbage collection
- O(1) crash recovery
This allows a unified architecture where KV, OLAP, vector, graph, and streaming workloads operate over the same epoch-versioned data model — not separate engines stitched together.
The implementation is heavily validated using Loom, Shuttle, Kani, and Stateright before anything ships. It's not a toy database.
Private for now. The open-source work above exists to support and harden this system.
None. I'm not currently looking for work.
Contact: [email protected]