A RAG-powered chatbot API that answers questions about a developer's professional background. I built it with Hono, LangChain, Mistral AI, and SQLite. I serve it on a Svelte site for my dev portfolio, but it works as a general API as well.
I dumped all my job background info into it: resumes, LinkedIn, cover letter snippets, GitHub repos, recommendations, projects, languages, volunteering, personal work, and even the story of building the thing itself.
It uses structured markdown files I created from all that as a RAG context. It doesn't automatically ingest LinkedIn or other services.
Manually. I went to my profile and saved 4 key pages as html. Then Claude converted those to markdown. Took less than 5 minutes. Automating that would have been more trouble than it's worth. The LinkedIn API is fairly useless unless you're in their partner program. Scraping could have gotten my account blocked because ToS violation. That core information rarely changes so it doesn't need to be dynamically imported anyway.
Are you saying that the partner program takes time and not worth integrating ? I was planning to gather data of users via LinkedIn oauth, haven't started it, but happy to hear your experience
It's more than time. That program requires explicit authorization from them. It's designed for larger corporate integrations, not developer experiments. I seriously I doubt I would get approved even if I went through all the paperwork. It's not just a matter of just signing up for an API key.
As for your plan, gathering user data is probably the main thing they want to prevent.
Location: San Francisco, CA
Remote: Yes, prefer on-site/hybrid in SF
Willing to relocate: No
Technologies: Svelte 5 & Sveltekit, React & React Native, LangChain, Typescript, Node.js, PHP, Python, Linux admin, WCAG, UX/UI, Writing & content creation
Résumé/CV: https://joshuacurry.dev
Email: see dev portfolio above
15+ years frontend and full-stack. Lots of experience at content companies. Teaching, writing, video, and DevRel experience. Leadership experience. Recently leveled-up with a bunch of AI middleware courses.
The maker movement evolved. It didn't disappear. Once the tools became accessible to a much wider audience, such as children, it became an integrated aspect of education. It also became a cultural tool. The author is focusing on a very narrow path to monetization and manufacturing. That wasn't the goal of the movement at all. That was how startup pitches tried to capture the movement and extract value. I see 3D printing machines that create structures out of adobe now. Huge ones. I see whole niche industries coming from laser cutters and CnC machines. People who started on Arduino boards now build music synthesizers and modular synth components. That movement continues and now offers a wide array of dividends.
This is a good tooling survey of the past year. I have been watching it as a developer re-entering the job market. The job descriptions closely parallel the timeline used in the post. That's bizarre to me because these approaches are changing so fast. I see jobs for "Skill and Langchain experts with production-grade 0>1 experience. Former founders preferred". That is an expertise that is just a few months old and startups are trying to build whole teams overnight with it. I'm sure January and February will have job postings for whatever gets released that week. It's all so many sand castles.
No idea about training tenserflow models - is it super complex or is it just calling a couple of APIs ? Langchain is literally calling an API. Maybe you need to get good with prompting or whatever, but I don't see where the complexity lies. Please let me know.
Having used both Tensorflow (though I expect they mean PyTorch which is way more popular, and I have also used) and langchain, they are nothing alike.
They he ML frameworks are much closer to implementing the mathematics of neural networks, with some abstractions but much closer to the linear algebra level. It requires an understanding of the underlying theory.
Langchain is a suite of convenience functions for composing prompts to LLMs. I wouldn’t consider there to be some real domain knowledge one would need to use it. There is a learning curve but it’s about learning the different components rather than learning a whole new academic discipline.
There's a big difference between building an ML framework like Tensorflow or PyTorch (I built a Lua Torch-like one in C++ myself) and just using it to build/train a model.
Building the model may range from very simple if you are just recreating a standard architecture, or be a research endeavor if you are designing something completely new.
The difficulty/complexity of then training the model depends on what it is. For something simple like a CNN for image recognition, it's really just a matter of selecting a few hyperparameters and letting it rip. At the other end of the spectrum you've got LLMs where training (and coping with instabilities) is something of a black art, with RL training completely different from pre-training, and there is also the issue of designing/discovering a pre/mid/post training curriculum.
But anyways, the actual training part can be very simple, not requiring too much knowledge of what's going on under the hood, depending on the model.
You're right, none of these new tools are disciplines. They are vendor specific approaches that are very recent. That's part of my overall point. Who is out there with 2+ years of very narrow tooling experience at another company at a senior level and is available for a rando startup (or desparate enterprise looking for bolt-on AI features) at a fraction of the pay? Not many, I'm sure. We can level up, do training, and maybe stand up a demo project. But that won't satisfy an ATS scan. It's unrealistic.
There are multiple services that verify and rate NGOs and nonprofits. The key is to look them up on the service website and not just Google the name. Personally I use Guidestar, but that's for U.S. orgs.
True. People define the speech of others as violence because they think it makes a violent response into self defense. It isn't true and never has been. If you're responding to speech with violence... you're the baddie.
I think packetlost knows that. I think the argument being put forward was that "if you are the kind of person that thinks that speech is violence, then you would believe that allowing someone a platform..."
reply