I am one half of DevRel (developer focussed aspects - node operation is maintained by the other half of the team) and do integrations for the mixnet project Nym.
A lot of my work now is on what DevRel looks like in the age of LLMs: docs and SDKs built so that an agent, and not only a person, can carry an integration through with as little friction as possible.
Current work centres around:
Recent work, newest first:
- A scenario suite for the docs-as-agent-surface, checking that an agent working only from the docs states Nym's limits instead of agreeing with whoever is asking
- An MCP server for the docs, so a coding agent can search the docs and the Rust source and read live network state; retrieval runs off a static index built at deploy time, with no vector database to operate
- Restructuring the network docs around an explicit threat model: adversaries split by vantage point, the linkage vectors that attribute traffic, and the unlinkability properties each configuration is measured against, with a decision path from use case to configuration
mix-*family of Typescript SDK libraries (drop-in-ishfetchandwebsocketMixnet abstractions for WebView-based apps + private DNS resolution)- crates.io preparation and publication flows for majority of the Nym monorepo crates
nym-smolmixRust crate, andnym-smol-core, the userspacesmoltcpstack underneath it- Rust SDK
streammodule (MixnetStream,AsyncRead/AsyncWriteover the mixnet) +IpMixStream, which layers the IP packet router on top of it to reach clearnet - Debugging WASM runtime errors and multiple (Go + Rust) runtime interaction for
mixFetch+ modifying its concurrency restrictions - Implementing the Diataxis framework for the Nym docs
Main focuses:
- System programming with Rust (especially focussed on tunneling/networking abstractions)
- Using WASM for privacy-preserving code within the confines of browsers (e.g. userspace TCP/IP, certs, and security restrictions)
- Documentation, including making it readable by agents as well as by people
Note on Git commit heatmap: since we use squash-and-merge for PRs, the majority of the commits in the heatmap below are merged PRs into the Nym monorepo.