Senior Software Engineer (AI + Agents)

We build the AI that major franchise headquarters use to coach every location in their network.

Think about a brand like a national fast food chain, swim school, gym, or wax studio. The headquarters doesn't own the hundreds—sometimes thousands—of locations operating under its name; independent owners do. But HQ is responsible for the performance of the whole network, and they steer it through a small field team of coaches, each accountable for dozens or hundreds of those locations.

It's a fascinating structure, and a hard one. Those coaches can't be everywhere, and they're buried in data flowing up from across the network—sales, customer reviews, audits, compliance records—with no clear answer to the only question that matters: what should this specific location do this week to improve? So coaching becomes educated guesswork.

Harmonyze is the platform franchise headquarters use to fix that. We're the performance coaching platform for franchise brands, and we sell to the brands themselves. Every day, our platform reads everything a brand tracks across its network and turns it into contextualized, prescriptive guidance: where each location should focus, why it matters, and exactly what to do next. It's working, network-wide, at headquarters across fast food, fitness, wellness, and home services, including brands like Goldfish Swim School, European Wax Center, The Picklr, and HomeFront Brands.

We're built in New York and backed by some of the most respected AI seed funds, alongside operators who know this industry cold. Franchising is more than $800B across hundreds of thousands of locations, and modern AI has barely touched it. We're a small team moving quickly to change that.

We're hiring a Senior Software Engineer to build the product at the center of it all—from the interfaces coaches and executives use every day, down through the data and the systems that turn a brand's complexity into guidance.

The role

This is a generalist engineering role, and we mean it. The hard problems here aren't only AI problems. We move a lot of data: integrations from a brand's many systems, ETL into a model that makes sense of it, database schemas that hold up as we scale, and the multi-tenant, enterprise-grade plumbing that keeps every brand's data isolated and correct. A good chunk of this is classic, well-built software engineering—and we care that it's done well.

The other side is the agentic system that produces the coaching itself: the orchestration that assembles context, calls tools across a brand's data, runs a multi-step analysis of a location, and returns something a coach can act on. We expect you to be comfortable here, not learning it from scratch.

And it's all in service of a product. You'll build the TypeScript and React interfaces people actually log into, and you'll care about whether a coach can act on what they see—not just whether the endpoint returns. The best engineers here think about the business and the user, not only the code.

What you'll do

• Build the product surface in TypeScript and React—the views where a coach sees their prioritized locations, understands why one is flagged, and gets clear next actions. Own components end to end and ship them to production.
• Build and run the data backbone. Integrations, ETL, and the database design that turns a brand's many systems into clean, queryable truth—and holds up under real volume across many tenants.
• Work on the agentic systems that generate guidance: context assembly, tool calling, multi-step orchestration, and the structured outputs the product reads.
• Keep it enterprise-grade. Data isolation, access, and correctness handled right, because brands trust us with their whole network's data.
• Own problems, not tickets. Early team, fast cycles—you'll have real say over what gets built and how.

What we're looking for

• You're a strong generalist engineer. You're comfortable across the stack and across problem types—frontend, backend, data—and you reach for the right tool rather than the one you know best.
• You've built and shipped product in TypeScript and React. User-facing features people rely on, where you owned the components, the state, and how it behaves under real data.
• You know how data systems actually work. ETL, relational database design and relationships, query performance, and the realities of moving messy data between systems. This is core to the job, not a footnote.
• You've shipped enterprise software, and you build with that lens. You understand what SOC 2, privacy, and customer data obligations mean in practice, and you design for isolation, access, and auditability from the start.
• You've deployed agentic systems in production, not just demos. You understand context assembly, tool/function calling, orchestration, and structured outputs—and the specific ways these systems fail and how to catch it.
• You're product- and business-minded. You can argue why something matters for a coach or for the brand, not just how it's wired.
• You've built real things at scale. Production systems with real users and real stakes—not a pile of weekend side projects. We're post-PMF and enterprise; you should be the kind of engineer that requires.
• You operate well in ambiguity. You fix the broken thing instead of waiting for someone to assign it.

Strong pluses

• Data science depth—causal inference, time-series, experimentation—anything that sharpens how we isolate what's actually driving a location's performance.
• Experience designing APIs or MCP servers that an agent, not a person, is the primary consumer of.
• You've been an early engineer at a company that scaled.

Why Harmonyze

• Your work makes people better at their jobs. The systems you build give coaches and operators a clear understanding of how to improve—and there's no ceiling on that.
• Foundational role. You'll be one of our earliest engineering hires and shape both the product and how we build it.
• A real product with real customers. Household-name brands already run their coaching on Harmonyze, and you'll build for problems that matter to their business.
• Hard, varied engineering. Frontend, data, enterprise systems, and AI—rarely the same problem two weeks running.

How we'll hire

• An informal conversation about what you've built and how you like to work.
• A practical exercise drawn from real Harmonyze problems—spanning product, data, and the systems behind the guidance—looking at how you think about reliability and code quality in a multi-tenant, enterprise context. No algorithm trivia.

Tech stack: TypeScript, Python, AWS, relational and time-series data stores, agentic frameworks and LLM tooling.

Come build the system that turns a brand's complexity into guidance their whole network depends on.

This is a NYC-based role with highly competitive cash and equity. Harmonyze is post-product-market-fit, well-capitalized, and working as a pioneering partner to household-name brands.