Location: San Francisco Bay Area (Hybrid)

About Carvis

Carvis is building the agentic operating system for the physical world of vehicle service and repair. Every day, millions of vehicles move through a complex ecosystem of service advisors, technicians, operators, and fleet managers — yet the systems supporting them haven't kept pace with the growing complexity of modern vehicles. Carvis transforms how decisions are made across the entire service lifecycle, turning fragmented data from real-world repair cases into instant, precise, and scalable intelligence for shops, fleets, and dealerships. This isn't just about efficiency — it's about redefining how work gets done in a $200B+ industry and laying the foundation for AI to operate in complex, real-world environments. We're a small, fast-moving team building at the frontier of physical AI, looking for people who want to shape the future.

The Role

We need a founding engineer who can own the AI systems at the core of our product. Not someone to slot into an existing stack — someone to define it. Here's a concrete example of the kind of problem you'd solve: A vehicle arrives at a shop with a check-engine light and a P0420 catalytic converter efficiency code. The system needs to pull that vehicle's full repair history, cross-reference common failure patterns for the make/model/mileage, retrieve current parts pricing from multiple suppliers, estimate labor time based on the shop's historical data, and surface a recommended repair plan to the service advisor — grounded, priced, and ready to present to the customer. All in seconds, and reliably enough that 500 shops can trust it every day. That touches data ingestion, retrieval, grounding, agent orchestration, and quality measurement. That's the job.

What you'll own

Agent orchestration & estimate generation — Design and improve the multi-agent workflows that transform a vehicle's symptoms, history, and parts availability into grounded, accurate estimates. You'll work across retrieval, context assembly, and generation. Evals, benchmarks & quality loops — Build the evaluation frameworks that measure estimate accuracy, grounding quality, and recommendation reliability. Estimate consistency is the core product promise — you'll own the systems that prove it works and catch when it doesn't. Data & retrieval infrastructure — Build and maintain the pipelines that ingest repair-order data from shop management systems, normalize messy real-world automotive data, and make it retrievable for downstream AI systems. Full-stack product delivery — Ship user-facing features end-to-end: front-end, back-end, and data layers. You'll work directly with founders and early customers to iterate on what advisors actually need. Technical direction — Help define the product roadmap, make architectural decisions, and shape engineering culture. As we grow, help recruit and mentor the engineering team.

What you bring

You've shipped AI systems to production in ambiguous, fast-moving environments — not just trained models, but built and maintained the full pipeline from data to user-facing output.
You have hands-on experience with agent architectures, RAG pipelines, or LLM-based systems — you understand retrieval, grounding, prompt engineering, and orchestration at a practical level.
You have strong data engineering instincts — you're excited about wrangling messy real-world datasets (think: inconsistent repair codes, free-text service descriptions, variable parts naming) and turning them into production-ready systems.
You've built evaluation and quality measurement into your AI systems — not as an afterthought, but as a core part of how you ship.
You can build end-to-end without a playbook — proven through startup experience, open-source work, or side projects that demonstrate full-stack ownership.

Nice to have

Experience with generative AI frameworks (LangChain, LlamaIndex, or similar).
MLOps pipeline experience — cleaning, labeling, and productionizing data and models.
DevOps chops — CI/CD, Docker, Kubernetes.
A genuine interest in cars and the automotive industry.
Strong communication skills and comfort collaborating directly with founders and early users.

Why Carvis

The market is massive and untouched by AI. Auto repair is a $300B+ industry in the U.S. alone, still running on phone calls, paper estimates, and tribal knowledge. The efficiency gap is enormous, and we're building the system to close it. Your work ships to real shops, immediately. This isn't a research project. Sun Auto's 500 locations and the 2,000+ shops in our pipeline mean your systems will be in production, handling real repair orders, from day one. You'll define the company's technical foundation. As a founding engineer, you won't inherit a stack — you'll build it. Your architectural decisions will shape how Carvis operates at scale. Compensation: We offer a competitive salary and meaningful equity package, with compensation determined based on experience, location, and role scope. Our goal is to ensure alignment between individual impact and company growth over time.

How to apply

If building the AI infrastructure for vehicle repair excites you, reach out. Send your resume to [email protected]. Carvis is an equal opportunity employer. We celebrate diversity and are committed to building an inclusive environment for all employees.

Founding AI Engineer