Call —
Transcript batch
0:00 Caller: Hey, how's it going?
0:00 You: This call will be recorded.
0:30 Caller: I'm well too.
0:30 You: I'm well and yourself.
0:34 Caller: Where about some city are you?
0:36 You: I'm in Long Island City, so just over the East River.
0:37 Caller: Oh yeah.
0:41 Caller: We're our offices in Soho, but it used to be Williamsburg, so next to close to LIC.
0:50 Caller: Some nice buildings in LIC, I feel like.
0:54 Caller: It's a lot of some new developments.
0:58 You: It's basically what?
0:59 Caller: Yeah.
1:00 You: But downtown Brooklyn was like maybe 20, 2008 or so, in terms of all the development and investment.
1:07 Caller: Nice.
1:08 Caller: I appreciate you getting back so quickly.
1:12 Caller: Plus a fast turnaround.
1:15 Caller: Yeah.
1:18 Caller: I mean, I think, yeah, the thing that caught my eyes,
1:22 Caller: I thought class looks actually pretty cool because we use a lot of cloud code,
1:26 Caller: but there's all these new models and you kind of want to
1:28 Caller: you kind of want to, I don't know,
1:30 Caller: think about routing them in a smart way
1:30 You: Sure. What would you like to go over?
1:32 Caller: or just like be able to play with them
1:34 Caller: in the hardness that you're used to
1:36 Caller: with all the context.
1:38 Caller: So that's why I was interested.
1:40 Caller: And then I saw you're also just generally like,
1:42 Caller: kind of have an interesting background,
1:44 Caller: having worked for a bunch of companies
1:46 Caller: and now at Roe for quite a while.
1:49 Caller: So I wanted to just reach out and see you.
1:52 Caller: I just talk to you basically.
1:58 Caller: Yeah, I mean, I guess first question for the class thing, is that, like, do you use that in your day-to-day or was that sort of like a side project?
2:00 You: I used to use it in my day-to-day.
2:12 You: More recently, I use something called needle.
2:15 You: Needle is what I consider to be the evolution of clasp, where clasp requires that you be present in the interactive CLI.
2:25 You: Needle is headless.
2:28 Caller: Yeah, interesting.
2:29 Caller: And do you run those cloud agents on some form of infrastructure, like cloud managed agents
2:30 You: Oh, no, I run my own infrastructure.
2:37 Caller: or, I don't know, ETV?
2:40 You: So I have remote dedicated servers and then a remote Kubernetes cluster.
2:42 Caller: Yeah.
2:45 You: The dedicated server is where most of my interactive work happens.
2:49 You: And then I dispatch the workers into Kubernetes to actually go forth and turn the specs or
2:54 You: thoughts or feedback into code.
2:58 Caller: Hmm, interesting. Do you use that at Roe or just in your private side-coding?
3:00 You: Both.
3:05 Caller: Okay. Nice.
3:07 Caller: And...
3:07 You: It's headless and it's independent.
3:09 You: So where I'm, prior to LLMs, I was productive at most six, seven hours a day actually writing
3:15 You: code.
3:16 You: Now it's closer to 24.
3:21 Caller: How do you review the code if you're generating this much?
3:21 You: I don't.
3:23 You: I look at the output of the application it produces.
3:26 You: And now if it's something that's
3:27 Caller: Okay.
3:28 Caller: Okay, nice. Well, at Roe, what are the things that you're working on there?
3:30 You: very sensitive or high or the risk of being wrong is considerable, then yeah, of course,
3:35 You: I'll review the code directly. But most of the time, it's downstream of the sensitive stuff.
3:48 Caller: Because I imagine that, like, all the finance and transaction stuff, probably you'll, people want to have a very close look at the code that changes that aspect of it.
3:57 You: Oh, I agree.
3:58 Caller: Um.
3:58 You: If it moves money, there is always going to be human review, because Roe is in the business of trust.
4:04 You: A misplaced penny is unforgivable.
4:06 You: But for everything else, operations, compliance, sales, internal finance, accounting, engineering, and so on,
4:15 You: there are plenty of things that can benefit from automating that work away and doing it in a way
4:22 You: way that is trusted and most importantly recoverable.
4:28 Caller: Yeah. Why did you choose to build your own info for that?
4:28 You: Well, all these features that we have today didn't exist when I first started doing it this way, or started using needle.
4:32 Caller: Was it just mainly the model routing or do you get any other benefits from running your own infrastructure over, say, using, you know, all the managed features from the cloud agents from cursor, from, from all these companies?
4:54 You: So I would argue they, they saw what I was doing.
4:56 Caller: Yeah.
4:58 Caller: Okay.
4:58 You: They found me at the top of whatever GitHub leaderboard and reverse engineered a good chunk of the stack.
5:04 You: Or more likely, a lot of people were doing the same ideas and implementing the same solutions in parallel
5:05 Caller: Okay.
5:06 Caller: Yeah.
5:12 You: because we're all using the same AI to come up with the answer.
5:19 Caller: Yeah.
5:20 Caller: I think everyone's kind of running into the same problems in the same sequence, essentially, and then solving for them.
5:27 Caller: solving for them. Nice. Well, I mean, I know you've worked at a lot of, like, big companies. You seem to be doing a lot of stuff on the, on the side. I think where we are, as I said in my message, I think, briefly, we're a YC company, and we're growing very quickly, and we're building agents and agent infrastructure for insurance. So there's a lot of these kinds of questions that come up for us as well, which is why I was interested in talking to you. And we're kind of building out
5:28 You: Thank you.
5:56 Caller: our engineering team right now. And so I just wanted to see how, if you've ever worked or considered
5:58 You: Part of the reason why I'm looking is because someone cold emailed me against my GitHub to suggest I interview at their company.
6:02 Caller: working for earlier stage companies, what you're, I mean, you seem to have been in a role for quite a while,
6:07 Caller: if you're starting to get bored or how you're just generally thinking of what's out there.
6:21 You: So that process is still live. And then anyone else who comes along, I'm happy to at
6:26 Caller: Okay. Nice. What was the, like, what's interesting for you? Like, what are the things that would make you consider joining a company?
6:28 You: that conversation.
6:40 You: So the three things I look for in any role. The first is data is both present and necessary.
6:46 You: Instinct and insight are good, but I'm of the opinion they must always be backed by data.
6:51 You: Second is a place where I have the relative freedom to pursue what I genuinely think is the best
6:56 You: solution to a problem.
6:56 Caller: Yeah.
6:58 You: My career is quite diverse and it's getting longer, so I want to be able to bring
7:03 You: forward the earned skills, knowledge, and experience into the next role.
7:03 Caller: Yeah.
7:04 Caller: Is any of those things not happening at Roe?
7:08 You: And the last is a place where the phrase, that's not my job, doesn't exist.
7:12 You: I see that as a sign of bureaucracy and I'm trying to avoid it.
7:20 Caller: Or is it more just like a curiosity of what's out there while you're sort of open to take calls?
7:25 Caller: Yes.
7:28 You: issue at Roe is I am the last New York-based engineer. There used to be eight of us.
7:37 You: And all of those, as engineers left, resigned or were laid off, they were replaced offshore.
7:37 Caller: Hmm.
7:44 You: So I imagine my time is coming soon, especially now that the initial grant has matured.
7:50 You: So I figure it's a good time to at least start looking and having those conversations.
7:55 Caller: Yeah. Nice. And I see that you, I mean, I know that neither is implemented in Rust. I see some go. I also see just generally a lot of like data engineering or like data science in your, in your background.
7:58 You: You know.
8:10 Caller: What's your current, like thing that you're interested? I mean, you said relative freedom is important, but like, what's the the tech stack that you'd be excited about working on and the kinds of problems you'd be excited about working on?
8:21 Caller: And then I can go a little bit into kind of what we're building and what would be helpful for us.
8:25 Caller: us. But I just wanted to see first kind of how you're thinking of software engineering versus data engineering versus a bit more research-y type things.
8:28 You: So I use code to solve problems.
8:37 You: It so happens that many of the problems I've solved in the past were predominantly shuffling data back and forth
8:42 You: or turning that data into insight.
8:45 You: In terms of the stack, if I'm allowed to agentically engineer it,
8:52 You: the stack is functionally irrelevant, only because the LLM would be the one writing it.
8:55 Caller: Thank you.
8:56 You: I focus a lot more.
8:58 You: more on the architecture and making sure what I build is something that can scale reasonably well
9:03 You: relative to however much it costs.
9:08 You: So planet scale is unlikely for most applications, but something that can survive multiple
9:14 You: availability zones going down.
9:16 You: That's roughly the level of resilience I try to build.
9:20 You: As for the problems I'm most adept at solving, it's the kind where workflows are predominantly
9:25 Caller: Thank you.
9:27 You: manual. So a classic example would be quite literally copying in data back and forth
9:28 Caller: Thank you.
9:29 Caller: Thank you.
9:34 You: across different spreadsheets. Knowledge of databases and SQL, the querying language, is not yet
9:43 You: widely distributed enough that being able to handwrite those pipelines is still a major advantage.
9:50 You: With the advent of AI and the different design patterns, the LLM-powered AI and the design patterns that come
9:56 You: with it. In theory, at least, the construction of these pipelines, these data
9:59 Caller: Thank you.
10:03 You: transformation should be easy. And watching firsthand, non-technical people attempt to build their
10:10 You: own dashboards in the upstream pipelines that feed it, there are definitely a lot more nuances
10:16 You: than simply knowing how to write the code. So AI took my job, absolutely, but to the next level.
10:23 You: So having seen, or at least I've dispatched my agents to get a better understanding of what,
10:29 Caller: Avalon.
10:30 You: is it a Valion?
10:31 You: I'm not sure how the correct, the correct pronunciation.
10:35 Caller: I mean, we'll focus on the claims process, I think insurance is quite a large space.
10:35 You: Atholome?
10:36 You: Okay, well, I had my agents look through it.
10:40 You: Is it exclusively the claims process, or is there more to it?
10:49 Caller: So there's definitely adjacent processes that get touched or are involved in the kinds of things we do.
10:53 You: Thank you.
10:57 Caller: But we're quite focused on.
10:58 Caller: focused on the claims portion of the insurance world because it's a very large space and it's like very regulated, very gnarly. And we thought if you have a lot of focus, you can get customers quickly and you can kind of build a book of business, which we did. And I think long term there is, you know, you kind of work towards becoming some form of like system of action, system of record where you then kind of funnel through.
11:23 You: Understood. And then what can you tell me about what can you tell me about what's running there about what's running there?
11:28 Caller: a lot of the operational processes that are happening in insurance companies. And once you've kind of exhausted the claim space, you can definitely go into the, like, I don't know, insured, like insured benefit side. You can go into the policy and underwriting side.
11:42 Caller: But we're focused on claims for now because it's a lot space.
11:53 You: I guess the gaps that you're hoping to fill.
11:58 Caller: high level, like what our product does is it's actually kind of similar to what you said, like
12:04 Caller: kind of replacing spreadsheets. We are essentially automating just like back office operational work
12:11 Caller: for these insurance companies in the claims domain across the kinds of things that the people,
12:18 Caller: the clerical people do, which is a lot of calls, emails, document processing, and then updating
12:23 You: Thank you.
12:24 Caller: kind of systems of record. And so our product is a fairly.
12:28 Caller: wide footprint to be able to cover that like full suite of back office work. But the core
12:35 Caller: of it, if you can think of it, it's basically an agent loop. You can think of something like
12:40 Caller: the entropic agent SDK that runs in a cloud container that has a list of like access to a bunch of
12:47 Caller: tools. So it can send emails. It has code tools. It can do browser scraping to update systems
12:52 Caller: of record, for example. And then it's handling various processes. So I don't know, you might have
12:53 You: Thank you.
12:57 Caller: someone submits a claim form, which is kind of like a scanned PDF from a messy insurance
13:01 Caller: form to a fax agent then reads it and parses out the details, creates a structured schema, puts it into
13:10 Caller: the system of record, and there's a bunch of data missing, and it dispatches some voice agents
13:14 Caller: that kind of call and follow up on it, and just kind of like works through, for example,
13:20 Caller: in this case, a claim intake, beginning to end, and just kind of like, yeah, either wakes itself up
13:23 You: So is it.
13:26 Caller: with crown jobs or gets woken up by various events like a call finishing, email coming in,
13:30 You: So is it the correct.
13:32 Caller: or these sorts of things. And I can speak a bit more about the tech tech, but I want to see if
13:33 You: The correct.
13:37 You: Is it the claims?
13:38 Caller: you, if that makes sense to you or if you've any questions on what the shape of the product broadly is.
13:51 You: Is it the claims processor?
13:52 You: Is it?
13:53 You: Or the adjuster?
13:56 Caller: is the term. It's not just the adjuster. Like, it's definitely different companies are organized
14:02 Caller: in different ways where the percentage of clerical work is, like, you know, distributed across,
14:08 Caller: you have claims adjusters, claims assistants, you have managers. Sometimes it goes even a bit
14:12 Caller: into like customer service, right, because people are calling in about their claims and what
14:16 Caller: information. So I wouldn't say it's just the adjuster, but you can think of it as concentric
14:22 Caller: circles around the job category of the adjuster. And then, like,
14:23 You: Thank you.
14:26 Caller: spreading outwards from that. And then different companies have, you know, sliced the task set in
14:31 Caller: different roles. But I would say we're definitely focused on just like the mundane, like, admin work.
14:39 Caller: And we're trying to be careful not to do too much adjudication ourselves, because then you get
14:45 Caller: into a bit more dicey territory. And there is so much, like, an absolutely mind-boggling amount
14:51 Caller: of just clerical work and just, like, really tedious work that.
14:53 You: Thank you.
14:56 Caller: people have to do in this industry, that this was one of the reasons that we got excited about
15:04 Caller: it was it's so boring that not that many people are excited about becoming adjusters. So there's a
15:12 Caller: huge labor shortage in the space where, I don't know, the median adjuster is like 55 years old,
15:18 Caller: and nobody is deciding to become an adjuster anymore. So these companies quite struggling,
15:23 You: Understood. So it's predominantly.
15:23 Caller: there's a lot of kind of market pull from,
15:26 Caller: them to find solutions. And I think AI in a way came just in time for a lot of these, a lot of this work.
15:36 You: So it's predominantly the different the different inputs, extracting the different inputs, extracting the necessary
15:42 You: out of it, into whatever standard
15:46 You: form the different insurance providers, with, I assume, with, I assume some sort of
15:53 You: original document.
15:56 Caller: like documents, but then you can think of some of the types of documents also being just like
16:00 Caller: emails or calls that with voice agents as well. And then the document that a call produces is
16:06 Caller: the transcript where there's information and kind of information provenance coming from that.
16:11 Caller: And then you have kind of an existing set of systems of record, which are always old and shitty
16:16 Caller: just like weird to integrate with where there's important context for the agents to pull or
16:23 You: Understood.
16:23 Caller: to like record back their work into.
16:26 Caller: Cool. And yeah, I think in terms of scale and just like you spoke a little bit about failover safety,
16:39 Caller: I think there is a, this is exactly the kind of moment we're in right now where we've grown very quickly.
16:45 Caller: So it's like three X quarter over quarter, but like at solid seven figures in revenue and we're signing
16:52 Caller: big enterprise customers now. So we're kind of feeding the pull. And then
16:53 You: Thank you.
16:55 You: Thank you.
16:56 Caller: are you know scrambling to make the the architecture of a platform scalable enough so we can
17:03 Caller: handle that kind of volume and those kinds of compliance requirements. So I don't know, for like an example
17:09 Caller: of the thing we did recently was to migrate from, you know, super base to just like AWS Aurora.
17:15 Caller: So you can have failover regions and then you like try to go down the stack of all the kinds of
17:20 Caller: like elements of a platform that need failover regions from the LLN traffic to just like,
17:25 You: I agree.
17:26 Caller: it's like basic infrastructure providers to the containers, to the database, to deployments.
17:31 Caller: There's just like a ton of, yeah, basically like enterprise readiness that we're building right now,
17:32 You: I agree.
17:39 Caller: which sounds like it would be interesting for you. Definitely something you've thought about before
17:43 Caller: and could have valuable input in.
17:51 Caller: Cool. Yeah, do you have any other questions? I can go a bit more into our tech
17:55 You: I guess, where do you see my
17:56 Caller: I can go into the kinds of customers we have. I'm not sure what's the most relevant for you. I can go into the team.
18:03 Caller: I mean, I think now that I know a bit more about you, I'm having more ideas than I had before.
18:07 You: is it to cover.
18:09 You: Is it to cover a gap?
18:11 You: Is it to cover a gap?
18:12 You: Is it to add more capacity?
18:15 You: Or is there it's more nuanced than that?
18:25 You: Thank you.
18:25 Caller: before. So my original intent was basically helping with the platform engineering side,
18:30 Caller: so making our platform scalable, robust, just like shipping code to the core platform.
18:38 Caller: But it sounds like you also have a lot of experience just like building agentic systems and
18:43 Caller: agenetic harnesses and applying that to, you know, operational processes. And so I think one thing
18:52 Caller: that would be interesting for us on that front is, you know, we're building these agents and agent
18:55 You: Thank you.
18:57 Caller: systems and we're solving problems and we're kind of building the harnesses as we go, but we don't
19:02 Caller: really have like a very strong opinion on it yet. So like basically like what are the types of things
19:07 Caller: that should be possible from an agent harness perspective? How do you architect the agentic piece of the
19:13 Caller: platform in a future-proof way and in a way where you can swap out different models and how do you
19:19 Caller: have, for example, e-valves, if there's lots of models, lots of agents running at scale,
19:24 Caller: like how do you figure out what's going on and how do you make that accessible to business people
19:25 You: Thank you.
19:28 Caller: that can't check every, like, trace, agent trace. So I think there's a lot of, in that direction,
19:35 Caller: a lot of researchy or, like, kind of designing questions that come up. Then I think internally,
19:42 Caller: we're also doing a lot of the things that you're doing. So we build our own kind of coding agent
19:46 Caller: that, you know, in the beginning was based on basically.
19:49 Caller: like an open claw that we then at some point deployed on AWS that as all the repos has a bunch
19:54 Caller: of MCPs to our data dog to our like meeting recordings with our customers to our internet
19:55 You: Thank you.
19:56 You: Thank you.
19:58 You: Thank you.
19:59 You: Thank you.
20:01 You: Thank you.
20:01 Caller: management so to like try to automate a lot of the software engineering like kind of building the
20:03 You: Thank you.
20:05 You: Thank you.
20:06 Caller: software engineering factory if you will and I think that that's that's definitely an area that that
20:07 You: Thank you.
20:09 You: Thank you.
20:10 Caller: that I see and then I think just generally you can apply these things to many operational parts of the
20:18 Caller: business, like hiring, like, I don't know, I think, yeah, hiring and kind of company building,
20:25 Caller: like internal operations, there's a lot of things that I could see being valuable to be automated
20:31 Caller: and that are sort of in that vein of data and processes being managed in spreadsheets currently.
20:37 Caller: But I think first order of business would be helping with, like, the software engineering
20:39 You: Okay.
20:41 Caller: and engineering core platform engineering team.
20:46 You: So what would the next
20:48 Caller: Well, if you're interested in interviewing, I think one of the next, like, we would, we would do a proper interview process.
20:49 You: what would the next steps be then?
20:58 Caller: So generally, an interview process is a first interview, which is a get to know, which I think we kind of did already.
21:05 Caller: We can also do another one if you want to just meet someone else for more just high-level chat and questions and just like background probing.
21:09 You: Thank you.
21:15 Caller: And then we do a technical interview, which is an hour long.
21:18 Caller: that is very much AI-enabled coding, sort of testing your ability to manage coding
21:26 Caller: and shift good code with them.
21:30 Caller: Like, we use a lot of TypeScript, so it would be kind of, I think, a TypeScript repo, but that
21:35 Caller: one is more the bit more grueling technical structured interview.
21:39 You: Okay.
21:40 Caller: And then we have a final interview, which is usually in person, making sure you've met everyone
21:43 You: Okay.
21:45 Caller: from the team and is more of a values focused interview and also one where you then
21:46 You: Sounds good.
21:50 Caller: have the opportunity to like in the office meet a bunch of people, questions about culture,
21:55 Caller: ways of working, these sorts of things.
22:00 You: Sounds good.
22:02 Caller: Okay. Would you be interested in kind of doing an interview? I think we could, because we've
22:08 Caller: already chatted now, I think I have a reasonably good idea to what you're about. Would you be interested
22:09 You: Certainly.
22:14 Caller: in doing the technical.
22:19 Caller: Okay, nice. I will then put you into like our hiring thing and we will send you an email to like
22:27 Caller: put your availability in of like when would work and then we can we can schedule it from there.
22:32 Caller: Maybe one more thing that I think is important for us is, it sounds like you've kind of
22:38 Caller: worked kind of semi remotely now when the team in New York was shrunk down.
22:39 You: Not a
22:44 Caller: are quite focused on building in-person culture, which sounds like it could also be
22:46 You: Not a
22:47 You: The only
22:48 You: the only
22:49 You: the only
22:51 Caller: down your alley if you miss the team aspect, but we are kind of an in-person company in-person
22:57 Caller: soho. And so I just wanted to put that out there and see if that works for you or if you
23:02 Caller: have any concerns about it.
23:05 You: the only push
23:06 You: I'd offer is the flexibility
23:07 You: to offer is the flexibility to attend to
23:08 You: to attend to my disabled wife and our autistic son when emergencies arise.
23:14 Caller: Yeah, that's very reasonable. So yeah, that's fine.
23:18 You: Some people don't agree with that, so I ask now instead of later.
23:22 Caller: Yeah, I mean, what's the, I mean, first of all, sorry to hear that. I mean, that sounds tough and so you're dealing with that.
23:23 You: It depends.
23:25 You: It depends.
23:31 Caller: What's the implications of that? Like, how often do sort of emergencies come up?
23:37 You: It depends.
23:38 You: for our kid, if he gets in trouble in daycare, I got to pick him up early to bring him home.
23:42 You: His mother cannot walk the quarter mile roughly to retrieve him herself.
23:44 Caller: Okay.
23:51 Caller: So do you think that's the kind of thing you're describing?
23:52 You: That's the kind of emergency that I anticipate needing my direct attention.
23:57 Caller: Okay.
24:02 Caller: Yeah, I think that's your work.
24:05 Caller: We can talk more about these sorts of things, I think, as we get deeper into it, but generally
24:08 You: Correct.
24:11 Caller: it sounds like you're up for in person with a caveat that.
24:14 Caller: there's personal things and personal priorities that will come up and and need flexibility around.
24:22 Caller: All right. Very cool. Then I think I will just kind of like follow up with an email. What's your email, by the way?
24:23 You: Jed at Jeddhardin.
24:36 You: I'll follow up in our LinkedIn.
24:37 You: so you have everything and then get started.
24:44 Caller: I just want to make sure that I see it on your stack already, RustGo Python TypeScript, that that TypeScript is fine and something you're comfortable with.
24:56 You: I agentically code in TypeScript.
24:58 You: I do not actually write TypeScript myself.
25:01 You: So if that is a requirement, I would be woefully underqualified.
25:07 You: I have applications built in TypeScript, but wholly because of agentic engineering.
25:14 Caller: I guess like, did you just get into the, yeah, actually, but you're, you would say in terms of like software engineering best practices, do you have a good sense of that or do you just sort of fully lean into the exponential and let the AI decide the architecture and these sorts of things?
25:37 You: I was a senior engineer at Meta prior to the advent of LLMs.
25:44 Caller: Yeah.
25:45 You: So at least when it comes to back-end stuff, not really concerned.
25:50 You: I've solved these kinds of problems that LLMs cannot get solved architecturally at planet scale.
25:51 Caller: Mm-hmm.
25:56 You: Front end, I'm a lot weaker.
25:58 You: So being able to get web pages loading faster is my general value ad, not necessarily making them prettier.
26:06 You: them prettier.
26:14 Caller: We'll see if the kinds of work that we do here and the requirements we have are like a match for what you bring and also what you're excited about, and then we can just assess from there.
26:28 Caller: And you'll also, that'll be with Fliander or more of one of the two is going to do that technical.
26:36 You: Thank you.
26:37 Caller: And you can, you can also kind of get deeper into that with that. All right. Well, I appreciate you.
26:44 Caller: getting back to me so quickly and excited for the next steps. Yeah, if you can just send me your email.
26:49 Caller: I noted it down. I just want to make sure that I got the right one in the LinkedIn message. And then
26:54 Caller: we'll, we'll be in touch and hopefully we can get something scheduled for next week. What's your
27:00 Caller: current availability? And I know you said you're in another process. What's your timeline looking like
27:06 You: What do you mean by timeline anticipated offers or what do you mean?
27:06 Caller: at the moment? Yeah, basically, like how far along are you in the
27:13 You: Some are advanced? Some are not.
27:14 Caller: and the other process, how quickly would you want to?
27:19 You: My objective is to have a signed offer before the end of July.
27:26 You: And I think I'm on track to hit that.
27:28 You: And then two weeks from there, start whatever role I end up starting.
27:33 You: So.
27:34 Caller: That's great. Okay. Nice. That's great. Maybe last question that I should just ask for the sake of complete.
27:44 Caller: you have a wife and kids and you've kind of had senior engineering roles before. Can you walk me
27:49 Caller: through how you're thinking of compensation, kind of salary versus equity, what your expectations
27:57 Caller: are and goals are as you evaluate different options?
28:03 You: Generally, I'm targeting on the order of $270,000 base.
28:07 You: Equity, I would need to diligence against the cap table and preference stack, whatever exists, especially if it's illiquid.
28:14 You: And then beyond that, I'm not really sure what other levers you have to pull.
28:14 Caller: Yeah, I mean, like I said, we have benefits as well. Like we all quite
28:28 Caller: sports focused or sort of health focus. So we have meals in the office that get paid for by the
28:33 You: Excellent.
28:36 Caller: company. And then we have a kind of gym slash health stipend, which most of us use for just, I don't
28:43 Caller: going to equinox or something because it's around the corner and making sure we stay in shape.
28:49 Caller: So I imagine those other sorts of things besides, you know, salary and equity that can be put on top.
28:57 Caller: And then health insurance is also 100% covered by the company.
29:03 You: And I guess the only other question I have is how did you find me?
29:06 Caller: Cool. All right. Well, yeah. I'm research. I'm research. I mean, I'm research. I mean,
29:13 Caller: there was also agents on my end that kind of went through to find, I'm not sure quite what
29:21 Caller: the prompt is, but basically I'm looking for senior engineers who have had some experience.
29:26 Caller: And then I, because it's quite hard to tell from the LinkedIn, I always try to direct my agents to
29:33 You: Thank you.
29:33 Caller: look at people's GitHubs and their blogs and just find something that's interesting that matches
29:38 Caller: kind of what we need. And then they propose candidates to me. And for you, the thing,
29:43 Caller: that I think caught my eye were your projects like needle and class because it's something that I had
29:48 Caller: thought about as well. So I was like, oh, that sounds interesting. I'm just going to kind of shoot you a message
29:52 Caller: and see if you're interested in chatting. I also like that you've done a lot of stuff in Go. I personally
29:57 Caller: really like Go. It's a great language. I like how simple it is. Or like clean it is. So I don't know.
30:03 You: Okay.
30:04 Caller: There was enough interesting things there. And I was like, all right, I'm just going to send you an email and see what's that.
30:10 You: Understood.
30:11 Caller: Or send you a LinkedIn message.
30:12 You: Well, you have a response now with my email.
30:13 Caller: Nice.
30:16 You: Are there, I guess, an expanded corpus of my web surface area.
30:21 You: So any follow-up questions, you can direct your agents to draw the answers from there.
30:27 You: And I am curious to know what those answers would be.
30:29 Caller: I will check it out.
30:31 You: Excellent.
30:34 Caller: All right, man. I appreciate you jumping on. And then I'll send an email with the follow-ups.
30:40 Caller: And then we'll just take it from there.
30:42 Caller: And hopefully we can do the.
30:43 Caller: us the next interview next week sometime. And then we'll see where that takes us and
30:50 Caller: we're at after that. All right. All right. Yeah, you too. Thank you. Bye-bye.
30:52 You: Thank you.
30:54 You: Enjoy the rest of your weekend.