Founding Engineer

• if you're delusional enough, never satisfied with your skill level, insanely hardworking, and want to carry a crazy amount of responsibility, where what you code on Monday impacts thousands of people on Tuesday, this is for you.

• you'll work on knowledge engines and multimodal agents for complex physical products, including expressive voice AI and codegen systems that explain what text alone can't, then lead the deployment of all of it to real customers.

• no intro calls. to get considered, you post your submission to our challenge. we review new submissions daily and respond to every one, so you'll hear from us very soon. from there, three more steps. pass them and you're in.

• $200k cash salary + 1% equity. We're backed by Paul Graham, Bloomberg Beta, Burst Capital, SV Angel, Weights & Biases, and more. Position is in person in SF, 6 days/week

Every product ever sold needs support at some point. That support falls into one of two buckets.

Bucket 1: Simple stuff. T-shirts, screen protectors, keyboards. You buy it, it shows up, maybe you ask “where's my order?” once. This is solved. Zendesk and a hundred other horizontal companies solved it.

Bucket 2: Hard stuff. $20,000 industrial heaters. HVAC systems. CNC machines. Car parts. Products where buying wrong means your building doesn't have heat or your manufacturing line is down. Support for these products can only be performed by highly trained domain specialists and there aren't enough of them.

If you're selling EV charging stations, your support person needs to be a certified electrician who understands local power grids, installation codes, and compatibility matrices. You can't hire this off the street. You can't outsource it overseas.

You'd think LLMs would have solved this by now. They haven't. Three years into the LLM era, penetration in this industry is very low & the reason is twofold.

First, off-the-shelf models don't actually understand these products. The knowledge lives in 48-page technical manuals buried on some manufacturer's website in terrible formatting — wiring schematics, compatibility matrices, installation diagrams that can only make sense visually. A general-purpose LLM can't draw you the diagram showing how to connect terminal A to terminal B. It doesn't have the spatial understanding or the product-specific reasoning to be a real technical advisor. So companies still rely entirely on human experts.

Second, even if the models were good enough, there are no harnesses to make them useful in the business.

No engine to capture deep technical knowledge about complex physical products and keep it updated. No way for a company to offload tribal knowledge from their senior technicians into a system. No way to see what questions customers are actually asking and feed that back into the knowledge base. No generative multimodal presentation and no expressive voice support.

Prox is building the best technical product expert for extremely complicated physical products.

A multimodal agent that can draw wiring schemes, share CAD models, process incoming videos from a technician in the field, and support people over the phone with voice that can pass the Turing test. To get there, we're solving multimodal knowledge graph building at a very deep level. A huge portion of your work will be developing SOTA knowledge engines that can truly understand complex physical products.

• We're always crystal clear with people. You can see it in this job posting — we post the exact numbers you're going to make (no ranges), what it takes to join Prox, and what you'll be working on. It's impossible to build a great company if you're not crystal clear about what you want from your people and where you're going.

• In the best spirit of YC, our north star is: are we building something people want? We can't afford to innovate for the sake of it. Solutions we're working on must impact thousands of real people almost immediately. The metrics we optimize are tickets processed, questions resolved, dollars generated for the company.

• The office must be filled with fun, enjoyment, inside jokes, and memes. You cannot outcompete people who are having fun.

• Knowledge Engine. You build the core engine that stores, relates, and reasons over product knowledge: manuals, compatibility matrices, installation guides, field observations, tribal knowledge. Deep infrastructure that the rest of the company builds on top of. Think Obsidian but for complex physical products.

• Emotionally Intelligent Voice AI. You're building phone support where a customer calls in, describes their problem, and gets walked through a solution by an AI that understands tone, frustration, urgency, expression. This means SIP trunking, telephony infrastructure, audio processing. You'll be building complex telephony systems, working with audio at the Hertz level, and figuring out how to make voice AI that people actually want to talk to.

• Code Generation as a Primitive. When something is too cognitively hard to explain in words, our agent draws it. Real-time diagrams, interactive schematics, visual walkthroughs generated through code. You take dense technical knowledge and figure out the most cognitively simple way to present it to a human so they can actually understand it and move on with their life. This leans hard on code generation tools and on your understanding of how people process information.

• Owning Customers. You will own customer contracts end-to-end. Demo to deployment to onboarding to ongoing. You'll talk to public companies, huge private companies, and lead pilots yourself. We give you real accounts from almost day one, first solo, then leading a team as we scale.

Step 1: Take-home challenge. This is your first point of contact with us. No intro call. You go to the challenge repo, clone it, and build a multimodal reasoning agent for the sample product. The agent needs to answer deep technical questions accurately, in a helpful tone, and not just in text. If a question is best answered with a diagram, it should draw one. If the answer relates to an image from the manual, it should surface that image. We will test it hard.

Step 2: Founder call. If your submission is great, the next step is a call with one of the founders. This is where you get to explain your solution and ask any questions about the company and your role.

Step 3: Technical interview. We're mainly testing graph theory to see how you think about representing knowledge as graphs. You will get a studypack and you'll have a few days to prepare. No leetcoding, just 60-90 min of problem solving with one of the founders. This part is meant to filter for raw intelligence and how well you break down complex problems into sub-problems.

Step 4: Work trial. You fly to SF to work with us for three days. We cover travel and stay. The expectation: ship something real to customers. We're going to see how you actually work.

Step 5: Offer. If all stars align, you get an offer.

You're probably early in your career. Maybe graduating, maybe an internship or two. We're young ourselves.

Here's a litmus test for whether this position is for you: you achieve something you've been striving for a long time. You finally get it. And literally one second later, it's not enough. Your brain immediately moves to the next thing. If there is never a moment where you feel done, this role is for you. There is no ceiling here.

We're building a team of Inglourious Basterds. People with crazy, delusional belief in their ability to learn, deliver, and face scrutiny. People who started with nothing and made something out of themselves. If that's you, apply.

You should want to be the best engineer ever. Literally. Prox cannot win if you're not the best engineer. That's why we'll do everything in our power for you to become exactly that. We're going to give you all the resources, support, and external pressure you need to become the best, to unlock parts of your imagination and skills you don't think you had and make Prox a great company. It's going to be one great, tough, and rewarding journey together.

—Dima & Greg, founders of Prox