Product Builder

Role: Product Builder

• Gurgaon, with US hours overlap and US dealer visits
• Full-time, Reports to the Head of Product

Hear the problem on Monday. Ship the fix on Tuesday.

Most companies split this job across 4 people. Support hears the complaint, a CSM writes it up, a PM prioritizes it, and an engineer builds it weeks later. By then the customer has stopped asking. We want 1 person who does all 4.

About Spyne

Spyne builds AI for US car dealerships. Our products run inside 1,500+ rooftops. Studio AI gets every car photographed, merchandised and online faster. Vini AI answers every sales and service call, text and chat a dealership gets, around the clock. Dealers buy outcomes: cars that sell sooner, phones that get answered, service bays that stay full.

As a Product Builder, you close the gap between what a dealer tells us and what our product does about it. You hear the problem, decide what to build, write the code and show the dealer the result.

What you'll do

• Own a set of dealer accounts end to end. You're their technical point of contact. When a GM, service manager or BDC lead has a problem, it comes to you.
• Diagnose before you build. A dealer says "Vini isn't booking enough appointments." You pull the call logs, find the intents it's mishandling, and work out whether the cause is a prompt, a workflow, an integration or a missing feature.
• Turn asks into product. You write the 1-page spec, decide what's a one-off config and what should become a feature for every dealer, and get it on the roadmap.
• Write the code yourself. Agent configs, conversation flows, CRM and DMS integrations, dealer-facing dashboards, eval sets. You ship to production in the areas you own and go through engineering review for core platform changes.
• Close the loop. You tell the dealer what changed, show the before and after, and check that it moved the number they cared about.
• Take your support shift. Part of every week is live support. It keeps you close to what's breaking.
• Find the patterns. When 5 dealers ask for the same thing in different words, you notice and turn it into 1 product change.

Who you are

• An engineer first. 1 to 3 years total, with at least 2 years shipping production code. You're fluent in Python or TypeScript, you can find your way around someone else's codebase on day 1, and you've built with LLMs: prompts, agents, tool calls, evals.
• A former founder, or close to it. You started something. Maybe it didn't work. You learned to talk to customers, decide with half the information and ship it yourself because nobody else would. We see that as the best preparation for this job.
• Someone customers trust. You can sit with a dealer principal who has no interest in AI and leave with the real problem. You explain trade-offs in plain English and you pick up the phone instead of hiding behind tickets.
• A product thinker. You know the difference between what a customer asks for and what they need. You can say no to a request and offer something better.
• High agency. You move without waiting for a spec, a ticket or permission. If something's broken for a dealer, you treat it as yours until it's fixed.

Bonus points

• You've worked in an industry and sold to non-technical operators (restaurants, clinics, home services).
• You've done a stint in support, solutions engineering, implementation or as a Forward Deployed Engineer.
• You've built voice or conversational AI, or integrated with a CRM or ERP.

Why this job

• You see the whole loop. You hear the problem, design the fix, write the code and watch the dealer's numbers move, often in the same week.
• It's the job we're building the company around. Every new Spyne product stays with the product team until its first 10 to 15 customers are live and succeeding. Product Builders are how that happens.
• You're early in a market that's moving. Anthropic, OpenAI, Decagon and Sierra all hire engineers to sit with customers. You'll do that work in a vertical where AI is still new and dealer feedback is direct.
• The path compounds. This role leads to Product Lead, Engineering Lead or running a business unit. Or to your next company, with far better odds than last time.

How we hire

5 rounds, each testing 1 part of the job. Rounds 1 to 3 run as a chain: the problem you uncover in round 1 is the one you design in round 2 and build in round 3.

• Round 1: Business and customer deep dive. We play a dealer with a real, messy problem. You ask the questions, find what's actually costing them money and play it back in their words. We're testing how quickly you get to the real problem and whether the customer would trust you.
• Round 2: Product thinking. You turn that problem into a product flow: who does what, where the agent steps in, what's a one-off config and what becomes a feature for every dealer. You write it as a 1-page spec and defend what you cut.
• Round 3: Build with agentic AI. You take your flow and ship a working change using the AI coding tools you'd use on the job, such as Claude Code or Cursor. We care about how you direct the agents, review their output and test what you ship, as much as the result.
• Round 4: Culture fit with the hiring manager. We talk through your story: what you built, what didn't work and what you'd do differently. We're looking for high agency and ownership.
• Round 5: HR and closure. Compensation, joining date and any open questions.