graerg

Hey! Former astronomer here; I used to work a lot with SDSS data products. Do you happen to have a link to any sort of photometry DB/catalog for Rubin? Basically I’m looking to grab a ton of data but I’ve been out of the field for a while and I’m not sure where they publish their data sets and which ones you’d recommend as being relatively accessible.

I built if-i-go-missing.com along these lines. Weird, I’m also a Brown!

Location: Orange County, CA (Pacific Time) Remote: Yes (preferred); hybrid Orange County also fine Willing to relocate: No Technologies: Python, Go, SQL | FastAPI, Django, Vue, HTMX | Postgres, TimescaleDB, DuckDB, Redis | Temporal, Airflow, NATS | AWS (S3/RDS/EC2/Redshift), Docker, Kubernetes | Prometheus/Grafana | LLMs in production (RAG, eval, orchestration, structured extraction) Resume/CV: https://brojonat.com/Jon_Brown_Resume.pdf Website / writing: https://brojonat.com GitHub: https://github.com/brojonat Email: [email protected]

PhD-trained engineer/manager (astronomy by training -- data is data). Currently Senior Data Science Manager at Hyundai Motor America's North American Safety Office, where I build large-scale vehicle telemetry pipelines, LLM-driven classification and RAG over enterprise data, and led department-wide LLM/AI adoption from technical workshops to production SDLC.

Previously managed a data science team at HeadSpin: built data science pipelines, killed error-prone spreadsheet workflows with proper APIs and dashboards, reimplemented the flagship data product to reduce costs, latency, and churn.

Looking for senior data / platform sorts of roles. Equally happy IC or manager -- I just want to ship. Open to full-time or fractional/contract; happy to even work on small projects you want to send my way to help get to know each other.

I build a lot on my own time, and the side projects are probably the best read on how I work:

- forohtoo -- Go service for awaiting Solana payments via SSE/NATS, with a clean `Await(memo, amount)` SDK. Writeup: https://brojonat.com/posts/forohtoo/ - IncentivizeThis (https://incentivizethis.com) -- bounty-based authentic ad platform using LLM-as-judge across Reddit, YouTube, Twitch, Instagram, HN, TripAdvisor. Writeup: https://brojonat.com/posts/incentivizethis/ - If I Go Missing (https://if-i-go-missing.com) -- dead-man's-switch service (Go + Temporal + Postgres + Twilio + Stripe). - IYBI ("If You Build It [They Will Come]", https://iybi-twc.com) -- websites for independent workers, managed entirely by email. Customer pays via Stripe, emails an AI agent (Claude + Cloudflare Email Worker + R2 + K8s CronJob), the agent researches them, builds the site, and handles ongoing updates. My own consulting site runs on it.

Mostly Go or Python services running on my personal k8s cluster. Happy to talk shop or small projects we could work on together to assess fit.

For my demos I've been running 3 minute sessions but I'm planning to run 30-60 minute sessions depending on the scenario. I want to see people push the boundary of what can be done in an hour (with an agent or without, it's up to the contestant!) and ultimately have the match VODs serve as entertainment but also reference examples for "good" development workflows.

I'm working on a competitive coding gameshow. I'm imagining a combination of great british bakeoff, battle bots, and dota. Basically contestants get dropped into a fully equipped dev machine (all the bells and whistles one could want/expect including neovim, agent harnesses, cool styling, etc and if you want you can always clone your dotfiles and stow them!). I've gotten a decent prototype that live streams from Fly.io sprites to twitch, and I'm able to voice over or have OpenAI do commentary on the match. I've got a demo here: https://www.twitch.tv/videos/2792893261. Still a ways to go, but it seemed like a fun way to tinker with Sprites.

It is already very plausible (and has been since the 1950s) without the advent of LLMs. This is just another layer on top of the preexisting and very plausible existential threats we already face.

Location: Orange County, CA (Pacific Time) Remote: Yes (preferred); hybrid Orange County also fine Willing to relocate: No Technologies: Python, Go, SQL | FastAPI, Django, Vue, HTMX | Postgres, TimescaleDB, DuckDB, Redis | Temporal, Airflow, NATS | AWS (S3/RDS/EC2/Redshift), Docker, Kubernetes | Prometheus/Grafana | LLMs in production (RAG, eval, orchestration, structured extraction) Resume/CV: https://brojonat.com/Jon_Brown_Resume.pdf Website / writing: https://brojonat.com GitHub: https://github.com/brojonat Email: [email protected]

PhD-trained engineer/manager (astronomy by training -- data is data). Currently Senior Data Science Manager at Hyundai Motor America's North American Safety Office, where I build large-scale vehicle telemetry pipelines, LLM-driven classification and RAG over enterprise data, and led department-wide LLM/AI adoption from technical workshops to production SDLC. Previously managed a data science team at HeadSpin: built data science pipelines, killed error-prone spreadsheet workflows with proper APIs and dashboards, reimplemented the flagship data product to reduce costs, latency, and churn.

Looking for senior data / platform sorts of roles. Equally happy IC or manager -- I just want to ship. Open to full-time or fractional/contract; happy to even work on small projects you want to send my way to help get to know each other.

I build a lot on my own time, and the side projects are probably the best read on how I work:

- forohtoo -- Go service for awaiting Solana payments via SSE/NATS, with a clean `Await(memo, amount)` SDK. Writeup: https://brojonat.com/posts/forohtoo/ - IncentivizeThis (https://incentivizethis.com) -- bounty-based authentic ad platform using LLM-as-judge across Reddit, YouTube, Twitch, Instagram, HN, TripAdvisor. Writeup: https://brojonat.com/posts/incentivizethis/ - If I Go Missing (https://if-i-go-missing.com) -- dead-man's-switch service (Go + Temporal + Postgres + Twilio + Stripe). - IYBI ("If You Build It [They Will Come]", https://iybi-twc.com) -- websites for independent workers, managed entirely by email. Customer pays via Stripe, emails an AI agent (Claude + Cloudflare Email Worker + R2 + K8s CronJob), the agent researches them, builds the site, and handles ongoing updates. My own consulting site runs on it.

Mostly Go or Python services running on my personal k8s cluster. Happy to talk shop or small projects we could work on together to assess fit.

That's the whole thesis; YAGNI.

I run my own temporal service in my k8s cluster; this setup is the backbone for almost all my applications. For simplicity, I opted for the postgres backend. You still need to run the 4 (?) other service (history, matching, frontend, ui, maybe others, definitely others if you want observability with prometheus/grafana, and tad bit more complexity if you want tailscale to get in there and poke around).

They ship Helm charts so reality is somewhere between "helm deploy" and "substantial ops burden". I don't have to touch it very frequently, but that is not to say I don't have to touch it. There's occasional releases and there have been times where (probably due to my inexperience with helm) I botched an upgrade and lost some data. And I've been on this journey for years; when I first started, they didn't have a Python SDK and it was one of my (many) excuses to learn Go. But anyway to your point, yes, if you're comfortable with k8s and Helm then you shouldn't have much of a problem running hundreds of thousands of workflows; if you want to really push the throughput and optimize cost you probably need to get creative the individual services and look into cassandra (maybe? idk).

They made a big point of explicitly advertising this as a feature with the GPT-5 rollout, no? Routing to cheaper models/less reasoning depending on the input prompt.

Exactly! It must be exhausting to have this huge preoccupation with determining if something has come from an LLM or not. Just judge the content on it's own merits! Just because an LLM was involved doesn't mean the underlying ideas are devoid of value. Conversely, the fact that an LLM wasn't involved doesn't mean the content is worth your time of day. It's annoying to read AI slop, but if you're spending more effort suspiciously squinting at it for LLM sign versus assessing the content itself, then you're doing yourself a disservice IMO.

You could make the case uv falls in this category (I just prefix all my pip commands with uv) though we have yet to see if astral will become a "successful business"; I'm hoping they pull it off.

Part of it is this tufte.css: https://edwardtufte.github.io/tufte-css/

There was a post on HN a few weeks ago about natalie.sh replacing blog comments with Bluesky threads. https://news.ycombinator.com/item?id=44826164

I was inspired by that and wrote up my experience integrating with my Hugo based blog here: https://brojonat.com/posts/bluesky-hugo-comments/

People keep citing this study (and it was on the top of HN for a day). But this claim falls flat when you find out that the test subjects had effectively no experience with LLM equipped editors and the 1-2 people in the study that actually did have experience with these tools showed a marked increase in productivity.

Like yeah, if you’ve only ever used an axe you probably don’t know the first thing about how to use a chainsaw, but if you know how to use a chainsaw you’re wiping the floor with the axe wielders. Wholeheartedly agree with the rest of your comment; even if you’re slow you lap everyone sitting on the couch.

I should have said “traditional ads” but yes you’re right. Product placement is one of many use cases. You could also sponsor bounties with more general requirements in order to drive engagement in a particular sector, or incentivize comments with a particular sentiment (e.g., I have a bounty for encouraging comments on hobbyist subreddits).

One very simple use case is making repetitive edits to a YAML (or similar) file. Sure, I can record a vim macro and/or try and conjure up some way to get it done with as few keystrokes as possible and hope I don’t fat finger anything along the way. Or I can just pipe it to llm and say [make this edit], and it just works.

> This is a significant advantage that an MCP (Multi-Component Pipeline) typically cannot offer

Oh god please no, we must stop this initialism. We've gone too far.

I forgot to note: I’m “dogfooding” this…there are bounties available for promoting IncentivizeThis, so have at them!!!

I’ve implemented Temporal activities that pull content for each supported platform/content type and then feed that into an LLM with the bounty requirements. It’s not very elegant but for an MVP it seems fine.

Now that tool calls and structured outputs are more baked into LLM APIs, I have some refactors planned, to make the Workflow more “agentic”, but I don’t expect the user experience to change all that much; the biggest improvement will be being able to dive deeper into the content, assess images, etc.

This has been my take, and maybe I'm missing something, but my thinking has been that in an ideal case there's an existing API with an OpenAPI spec you can just wrap with your FastMCP instantiation. This seemed neat, but while I was trying to do authenticated requests and tinkering with it with Goose I ended up just having Goose do curl commands against the existing API routes and I suspect with a sufficiently well documented OpenAPI spec, isn't MCP kinda moot?

On the other hand, in the absence of an existing API, you can implement your MCP server to just [do the thing] itself, and maybe that's where the author sees things trending.

Not in my experience; I think I gave up and opted for mitmproxy which works but is not this easy/seamless.

I totally agree. My counterpoint is that this tool is the equivalent of portable stair lift. Yes, it may hinder me from developing the ability to walk up the stairs, but I'm fine with that. The only thing that matters is that I get up the stairs.

>A Sysadmin would take 5 minutes with a shitty language and a shitty tool and get way more done in less time.

Most people aren't sysadmins, but occasionally have to do sysadmin-like things. I've been programming with python and go for years. I've never been able to get the "core" command line utilities to really stick in my head. A sysadmin uses them every day, whereas I rarely have to reach for them. On the rare occasion when I _do_ have to reach for them, it is excruciating (what was the flag I need for `find` again?). If it were life or death and I had to debug even the simplest sed/awk command, it would be death for me! But this package makes perfect sense to me and really enables me to write this sort of quick and dirty thing in a language I'm familiar with and can confidently maintain.

This isn't for everyone, but there's definitely a population that can get a lot of value out of this.

In this vein, do you know what the experience has been like for users that use the vim vscode plugin? I see that it's trivial to carry over extensions and whatnot, but is the experience still "good" trying to use both?

Location: Orange County, CA

Remote: OK

Willing to Relocate: No

Technologies: Python, Golang, JavaScript, SQL

Resume/CV: https://brojonat.com/doc/Jon_Brown_Resume.pdf

Email: [email protected]

Astronomy PhD turned full stack data scientist. I've been working as a data scientist in the industry for just under 5 years. I like building HTTP services and also have experience collecting, cleaning, and modeling data.

Honestly that doesn't seem _that_ bad to me. It's a reasonable way of saying "hey, you're going off the happy path for this product and doing so at your own risk". This is an interesting gray area of consumer/hacker technology where bad software can destroy your hardware; there's probably difficulties in supporting warranty related breaks. Is the alternative to just replace parts every time someone bricks their hardware regardless of any/all modifications they've made? Maybe that's what they should do, but I can appreciate that that's a pretty tall ask, especially when the industry is still in it's infancy.