If you're looking for a job in early stage tech,

If you're looking for a job in early stage tech, here are 10 series B or younger startups that I think are doing some really cool things and are worth considering. On the AI coding side, you have two AI coding review startups, greptile and coderabbit. Coderabbit is Series B, they raised 60 million, and then greptile is Series A, they raised 25 million. That I think are both leading the pack. I would say that they're two of the most, if not the most widely used AIP review or code review tools out there. What I think is so interesting about them is that their product right now is such a great wedge into building a bigger platform around the software development life cycle. Also in the development life cycle, but in a different lens, is this company called Tessl. This is founded by one of the founders of Sneak, the security open source company. They essentially have a repository on steroids for agents, so they help you create agent skills, distribute them, have enterprise grade sort of compliance and workflows around them. So for a lot of people are using AI agents, you're definitely using skills and they can be powerful. And so this company really helps you supercharge them. This is a really interesting wedge into just the general software development life cycle. And from there, you can really start to expand, obviously the founder having done this at Sneak before and Sneak being arguably very successful, um, is a is a strong founder to sort of like back on the. On the search side of things, you have two companies that are doing similar things that are competitors. You have Parallel, which is founded by Prague, who is the former CEO of, if you followed Twitter, if you followed any of the Twitter versus Elon stuff, they just raised $100 million Series B. And then you have Exa, which does something similar. They both do real time, essentially a generic search at the scale of the web. Although to be fair, I think Exa just raised a big series C, uh, fundraising, so maybe a bit older than Series B, but it just happened in. Anyways, the main thesis here is that we as humans use the web in a certain interfaces, um, but agents have a different way of using the web. Um, there's almost this sort of like shadow internet that needs to be built so that agents can use it successfully. Um, because they don't care about traversing, they just wanna see, uh, you know, specific, sort of like call to actions that actually resonate and have certain, uh, meditated embedded inside of it that we humans don't really care about. And so at the end of the day, I think companies can grow really start with search. Obviously, you have Google, which was a search company, is a search company in a lot of ways. So I would really consider one, I really like Exa, um, and I really like Parallel as well. And I do a lot of search at work. So this, uh, this is something I'm interested in. There are two AI infrastructure companies that are pretty interesting. They're more on the model routing side. I think this is, I think this is interesting. One is obviously OpenRouter, the leader in, I would argue, in the space. They just raised $113 billion recently for Series B. And then you have a smaller company, I believe they're YC, if I remember correctly, they're Series A. They raised, I think, $20 million. I think model routing is becoming, I think it's one of those overlooked aspects of what a lot of teams are doing, but not everybody really talks about it, which is routing requests across many, many models to try to find the cheapest one, one for the task, the one that may have the least amount of latency or whatever, whatever other, like, sort of metrics or capabilities that, any one particular team requires. OpenRouter, I think is really good at that. I think Vellum can also be really good at that. What's fascinating is that I think this is gonna grow in the future because, you know, not only have, obviously, Chat GPT and, uh, and Claude, um, but then you have all the other, for example, Chinese models like Deep Seek, um, Kimmy, Quen and so on. Um, Mistral from France, who knows what they're doing, but it is a model that's, um, and what's fascinating is that building this sort of middleware infrastructure to be able to intelligently decide which prompt goes where based on how you wanna optimize one aspect of that, whether it's cost, latency, uh, intelligence, response, whatever. I think is something that is only gonna continue to grow, especially if you look at, um, uh, models or, uh, Fireworks or others. There are heavy, there's some really interesting interviews I've, so if you're considering these companies, um, I would look seriously into. I think they're actually pretty. Security side of things, are two companies that are doing some pretty interesting, but in different ways. You have a regular, a Series A. They raised, I think about $80 million. They're more like an AI research lab for security. So they publish a lot of research, they do research on things like model evals, things like model misuse, alignment, model safety and so on. Um, a lot of cyber stuff. Um, super interesting because they're one of the only AI sort of security research labs dedicated to security. Everybody kind of does it a little bit. Um, but super cool. I actually reached out one point last year to learn more about it. Um, got contact with anybody, um, super cool. Definitely something I would look into. The second one is Arcade. I actually gave a talk there, uh, recently. This was two weeks ago before I went to Japan. They're essentially building this authorization malware layer for agents and for MPC servers. So making sure that the agents, the servers that you're actually have the right authorization and authentication to be able to call tools that they wanna call. Um, and being able to enforce those and audit trail those. Authorization authentication is always needed. I mean, how many off and are there? Um, they're kind of like always. Yeah, they do reasonably well because every single company needs it. Um, I think the founder used, built a previous authentication company and sold it to Octa. Um, so he's obviously well known in the space. The thing I'm gonna do in the science space, which is Periodic Labs. They're actually fairly well known at this point, but they're still their Series A, they raised a big $300 million Series A around. They are funded by a number of people from Google, as well as one of the creators of Chat, GPT. I just generally love the idea of automating a lot of science discoveries, being able to take in different, different experiments, different data from all these different sources, to iterate that, um, automatically using AI, using of like a, is very obviously gonna be something in the future. I expect that probably another opening eye or anthropic, uh, company to be built, particularly within that space. I think they have a good shot of doing it. It's something that like I'm more and more getting interested in, and I would strongly consider.