Write the hot path. Decoders, streams and transaction routing where microseconds show up in the product and in customer results.

NoLimitNodes parses over a thousand typed events from thirty-seven Solana programs in real time. HyperliquidRPC serves RPC with a 4.2ms median and archives hundreds of millions of events a day. The code that does this is mostly Rust, and it is the part of the stack where performance work pays off immediately.

This role sits in that hot path. You will write and tune decoders, stream processors and routing logic, measure everything, and watch your improvements show up in customer-facing latency numbers the same week.

What you will do

- Build and maintain high-throughput decoders and stream processors in Rust
- Profile and optimise the hot path: allocation, serialisation, network and syscall behaviour
- Keep parsers current as on-chain programs upgrade and layouts change
- Design schemas and protocols that stay stable while the data underneath moves
- Build backpressure, replay and recovery mechanics that survive real traffic
- Own benchmarks and regression tests so performance does not rot quietly
- Work with infrastructure engineers on placement, networking and capacity

What we are looking for

- Solid Rust, or deep systems experience in C or C++ with real interest in moving to Rust




- Experience with concurrent, latency-sensitive or high-throughput systems
- Comfort with profilers, flame graphs and reading what the machine is actually doing
- Good instincts for binary formats, serialisation and protocol design
- Honest engineering habits: benchmarks before claims, tests before refactors
- Able to work from our Toronto office part of the week

Nice to have

- Familiarity with Solana internals: accounts, instructions, Geyser, Yellowstone gRPC
- Experience in trading systems, market data or MEV-adjacent work
- Time spent with io_uring, zero-copy techniques or kernel-bypass networking
- Experience running the things you build, not just writing them

Who you are

- You find a 20 percent latency win genuinely exciting
- You measure first and optimise second
- You read other people's code to learn, not just to review
- You want your work to have a number attached to it
- You are comfortable being on call for the systems you build

How we hire

- 1 Intro call with an engineer, about 30 minutes
- 2 Short take-home assignment involving real decoded Solana data
- 3 Technical conversation about your assignment, performance work and past systems
- 4 Conversation with the founders
- 5 Offer

#J-18808-Ljbffr

📌 Rust Engineer, Trading Infrastructure (Toronto)
🏢 CLR3
📍 Toronto