- Own the architecture of major subsystems — a new integration category, sync engine internals, our AI product infrastructure. The design decisions are yours to make and defend.
- Set the patterns the team builds on. Your abstractions become how we build; you influence through design docs, reviews, and tooling.
- Scope and de-risk multi-week projects yourself: identify the unknowns, prototype early, sequence the work.
- Make pragmatic build-vs-buy and tech-debt tradeoffs, and articulate the reasoning clearly.
- Raise the team's technical bar through design reviews, pairing, and mentoring the engineers we hire alongside you.
INTERESTING TECHNICAL CHALLENGES?
AI & agents: this is where most of our new product work is. We're building agents that read, reconcile, and act on construction financial data (invoices, change orders, pay apps, budgets) inside systems that were never designed to be queried. The hard parts are grounding, getting an agent to be right about a customer's actual numbers, evaluating quality when the correct answer lives in a 30-year-old ERP, designing tools over our own APIs, and deciding what should be a deterministic pipeline versus what to leave to a model.
Unification: we unify dozens-to-hundreds of fragmented systems under a single standard that we define: data objects, data models, authentication, account-linking UX, and features like filters and pagination. It's the foundation the AI products sit on, and every new feature has to be designed generically across every system we support and plan to support.
On-prem: we're connecting with many types of on-prem systems (SQL-based, DLL-based, API-based), many built 20-40 years ago. We provide tremendous value to customers allowing them to interface with these systems in a secure, fast, reliable way, as if they're modern, cloud-based products. We need to support real-time communication and webhooks when they don't exist (e.g., using web-sockets).
Scale & speed: we're handling millions of API requests per day, and growing daily. For our products that we power with our own API (e.g. Analytics), we need to store and retrieve large amounts of data in an efficient way, building elegant caching layers to reduce latency. We ship code daily and iterate quickly based on actual feedback from paying customers (they share tons!).
TECH STACK?