Forward Deployed AI engineer

Applied AI Engineer

SoHo, New York City · On-site · Full-time

$185K–$325K base + competitive equity, benefits, relocation assistance, and visa sponsorship

The company

This is a direct partnership with General Catalyst built to transform critical institutions with applied AI.

The team embeds with customers in healthcare, manufacturing, energy, banking, and similar environments where software has to work inside real operations, not just in demos.

One example of the kind of work this unlocks is General Catalyst’s $2 billion hospital system in Ohio.

The team pairs forward-deployed engineering with embedded product managers and researchers, and has strategic relationships with Anthropic, McKinsey, AWS, and companies within the General Catalyst portfolio.

The company was founded in 2025 , has 55 employees , and is planning to scale from 30 to 100 next year.

The role

This is a customer-facing applied AI engineering role for someone who can move from executive-level problem framing to production software.

You will work directly with customer executives, operators, and technical teams to define strategy, choose high-leverage use cases, build the systems, and then turn the lessons into reusable product capabilities.

You should expect meaningful customer time, but the output is software: decisions, systems, and product primitives that survive repeated use.

This is not a prototype-only role. The job is to ship production AI systems that work with real users, real data, and real operational constraints.

The technical problem

Critical institutions do not need generic chatbots.

They need AI systems that can sit inside existing workflows, respect organizational constraints, and remain useful when the inputs are messy, the stakes are high, and there are multiple human handoffs.

The hard part is translating frontier model capability into systems that operators trust and executives are willing to expand.

The company’s edge is that implementation and productization happen in the same loop: what works in a customer engagement should become a pattern in Mosaic, the in-house toolkit for rapidly deploying agentic workflows.

What you’ll own

• AI strategy with customers: partner with executives and operators to define the roadmap, scope the transformation, and identify where AI can remove bottlenecks or unlock new workflows.

• Use-case selection: work backwards from business objectives to prioritize the few problems worth solving first.

• Production systems: build and ship AI agents and workflows that work in the wild with real users, real data, and real constraints.

• Productization: encode what you learn into Mosaic through features, patterns, and abstractions that make the next deployment faster and more reliable.

• Cross-functional execution: collaborate closely with embedded product managers, researchers, and other engineers to align model behavior, workflow design, and deployment approach.

• Engineering judgment: make tradeoffs explicit around reliability, maintainability, latency, and operational simplicity.

Who this is for

You are likely a strong fit if you have:

• 3+ years in AI-focused software engineering, with the real bar being production ownership rather than years alone.

• Shipped production-grade AI agents or workflows, ideally including LLM-powered systems.

• Hands-on experience with LLM tooling and a clear sense of where model capability ends and software design starts.

• Worked directly with customer stakeholders or internal users to turn ambiguous requirements into deployed systems.

• Strong product judgment: you know which problems are worth solving, which are not, and what needs a model versus workflow logic versus human review.

• Comfort operating in high-stakes environments where the cost of a bad deployment is real and the quality bar is high.

• The ability to explain technical decisions to executives without hand-waving.

• Enough technical depth to be hands-on, and enough scope awareness to own the full arc from discovery to deployment to iteration.

What will separate strong candidates

• You have built something that went from a promising demo to a system people relied on.

• You are comfortable being the engineer in the room with a customer’s executives and operators.

• You care about whether the system survives contact with real users, not just whether the model output looks good in a notebook.

• You can move quickly witho…