Services

Embedded Rust expert

You already run Rust in production, or you are about to write some, and the expertise is missing on the team. I join your engineers to design, code and ship.

• ▸ Rust backend development on AWS: Lambda, ECS/EKS or EC2, depending on what the workload actually calls for
• ▸ Infrastructure as Code, CI/CD, observability and comments shipped with the code, not six months later
• ▸ Taking over an existing Rust codebase: technical debt, version upgrades, a CI pipeline put back on its feet
• ▸ Code review, and your developers getting better at Rust along the way

I work exclusively on AWS. You could object that it is narrow, but it is exactly what lets me stay in control of the project, from the architecture diagram down to the production code.

Architecture and performance

Nine years of AWS architecture and more than three years of Rust. When something costs too much, answers too slowly, or scares everyone away from touching it, I diagnose it and I fix it.

• ▸ Cold starts, p99 latency, a Lambda bill going the wrong way: I measure first, I put numbers on it, and then I propose a plan. Sometimes the answer is Rust, sometimes it is not
• ▸ Rust code review: correctness, idioms, and above all maintainability by your own team
• ▸ Serverless architecture review: API Gateway, Lambda, AppSync, DynamoDB, EventBridge, Step Functions
• ▸ An argued service choice between Lambda, ECS/EKS and EC2, with the 12-month cost set right next to it

What you get back: a short, prioritized, quantified report, with the measurement conditions. But I am not a generic AWS architect: when the plan involves Rust, I stay and implement it.

Open source crate support

Three crates I maintain exist to run Rust on AWS: lambda-appsync, dynamodb-facade and awssdk-instrumentation. If your team depends on one of them, you can have the maintainer's direct line.

• ▸ Support subscription: a contract with a day pool for incidents, questions and hands-on help
• ▸ Priority fixes: whatever blocks your production goes to the front of the queue
• ▸ Sponsored feature development: you fund exactly what you are missing
• ▸ Hands-on help integrating a crate into your codebase, and a live walkthrough for your team

I also develop new crates under sponsorship, given a clear use case and a dedicated budget.

Rust interests you, but you are not going to commit a team on a bet. Fair enough. So we start with the only question that matters: is it worth it, in your shop, and on which workload? And if the answer is no, I will tell you so.

• ▸ Audit of your workloads: which ones genuinely benefit, and which ones will gain nothing
• ▸ Expected gains put in figures (latency, memory, AWS bill) with the cost of the migration set right next to them
• ▸ Proof of concept on a real workload: numbers measured in your environment, not blog benchmarks
• ▸ Phased roadmap: what migrates, in which order, and with which safety net
• ▸ Your developers trained along the way, because the code I deliver has to stay maintainable without me
• ▸ Presentation to decision-makers: I take the demo in front of your leadership