Ask HN: Who wants to be hired? (June 2026)

Remote: Open to

Willing to relocate: Yes

Technologies:

* AI/ML- AWS Strands, Langchain, Langgraph, Vercel AI SDK. Agentic workflows and orchestration

* Web - Rect / Next.js / vercel

* Embedded - C/C++

* Infra - Extremely fluent in AWS / terraform / CDK

* Other - Java, Swift, JS, Node, Python

See resume for additional details

Résumé/CV: gasek.net/resume

Email: hnhired@gasek.net

Highly autonomous, motivated engineer and founder, seeking opportunity for broad impact and ownership and growth potential

Freelance or full time.

Remote: Yes

(or in person in west end Toronto or Oakville, Burlington, Hamilton, or Kitchener-Waterloo)

Willing to Relocate: No

Technologies: Rust, C++. Systems-level development, database internals, data structures, embedded systems

Summary: 25ish-years-experience systems software engineer. 10 years at Google. "Systems" level programmer, embedded systems, data structures, database internals, mostly in Rust. Plenty of open source experience. Looking for freelance, greenfield and/or interesting projects. Recent work all in Rust.

https://github.com/rdaum/ https://codeberg.org/ryandaum

https://timbran.ca/

Email: ryan.daum@gmail.com

  Location: Victoria, BC, Canada
  Remote: Flexible
  Willing to relocate: No
  Technologies: TypeScript, Node.js, Deno, Go, SQL (mostly psql, duckdb, sqlite), React, Vue, Docker, Full-stack, AWS
  Résumé/CV: https://steve-adams.me/resume
  Email: steve@steve-adams.me

Lately I'm in the spaces between data science and engineering reliable, high-performance, resilient, yet highly-usable tools for scientists and other developers, across the stack. It's a lot of fun.

Some recent projects:

- Performance tuning and infrastructure cost-reductions on https://oceanconnect.ca. It's faster and cheaper to run than ever. Costs are essentially cut in half, and performance across hot paths, on average, is about 30% better. There's now comprehensive observability across the stack, making it so we can rapidly diagnose issues with complex model processing pipelines and data acquisition systems. On top of that, the error rate has dropped dramatically despite increased visibility into what's going wrong (a lot of issues were silently swallowed by OOM culling in Lambda, for example). Also, I managed to reduce the storage footprint of the database by around 80% by more aggressively pruning obsolete data (especially raster tiles), tightening the sliding window around useful and aged-out data. This was done in part by partitioning data better so we could free it sooner without worrying about dead tuples and autovacuum. It's tighter, cleaner, faster, and cheaper to back up.

- Developing a low-cost solution to autonomously regulate temperatures in 40 saltwater tanks in a remote lab, using equipment which doesn't support remote controls. Using digital potentiometers inserted into each temperature controller, the system spoofs the tank's temperature to trigger heating or cooling, then observes until the correct temperature is reached. It uses several dissimilar redundancies, including a dead man's switch, to ensure there can't be run-away heating or cooling loops. The cost for new hardware with remote control capability came in close to $100K, but this solution (along with supporting software for designing temperature profile time series, monitoring tanks, and viewing historical data) is under $15K, yet required much of the same supporting software work. It's fully compatible with future hardware supporting direct remote control, as well; we'd only need to remove the temperature-spoofing loop and send set points to the units instead.

- A library for analyzing biodiversity datasets, inferring semantics about their data, and then creating a declarative configuration which can be used to transform the data to Darwin Core and validate it. It has a core library that can be used in any kind of client, like a CLI, web API, or a GUI. The declarative configuration makes it so datasets can be worked on in GitHub, then the CLI can run in a workflow to ensure it's still valid as it changes and grows. It uses DuckDB to make assertions about primitive types and data/relational integrity, then uses the application layer to make semantic assertions according to the specification's requirements. The entire workflow is highly supportive to scientists, explaining any issues with the data in very plain terms from start to finish. This has made it so scientists can make old records conform to the standard and have it ready for publishing far faster and with higher confidence than ever. At the same time, it also serves the purpose of helping scientists understand a large, complex standard one field at a time as they work through issues, rather than forcing them to tackle the entire standard from the outset. Ideally, this will lead to them working with the standard earlier and more often, with better knowledge and intuition for how to do so. It's built using Deno, Effect, Qwen 3.5 (for initial semantic inference passes), and DuckDB.

I'm also interested in LLM harness development and creating more deterministic agents with safety and behaviour guarantees.

My work experience is diverse, ranging from SaaS to science to a bit of firmware and robots. I'm interested in just about anything so long as it's genuinely useful, challenging, and the team is excited.