Founding Sales Engineer - Agentic AI Startup

FOUNDING SALES ENGINEER needed at one of our Agentic AI STARTUP clients!!! Hybrid 3 - 4x per week in office in NYC only. Base $200k - $230k plus around 20% target bonus and EQUITY. Looking for 5 - 12 years of Enterprise SE experience within firms that sell to insurance, asset management or private credit. Must have strong understanding/experience with Applied AI and ideally some startup experience.

Our client is the agentic workforce for regulated finance. They build AI agents trained on how a specific firm actually works — its operating logic, its edge cases, its regulatory framework — that run inside the client’s own cloud and governance boundary. Every agent they build belongs to the client: their cloud, their model, their IP. Outputs are auditable, traceable, and reversible.

They help leading institutions rethink critical workflows across insurance, private credit, private equity, and commercial lending. In insurance specifically, that means advisory, claims, financial planning, and the investment office.

The Role:

You are their first sales engineer. Today their CTO and founders sit in every technical conversation, which does not scale and is the reason this seat exists.

You will be the technical partner to their Client Partners and Account Executives across the full cycle: discovery, demo, architecture and security review, workflow scoping, and the handoff into delivery. You will translate in both directions — turning a client’s operating problem into something their engineers can scope, and turning what they have built into something a business buyer understands.

Be clear-eyed about what this is. No team, no library of demo assets, no discovery framework waiting for you. You will build the technical sales function while carrying it.

The problem you are solving:

Their demos currently run too technical for business buyers. The people who write the check are a claims operations leader or a head of financial planning, not an engineer. They need to see their workflow, their exception cases, and their audit trail — not a model architecture diagram.

At the same time you have to hold your own with a CIO, a model risk committee, and an InfoSec team, and answer questions about lineage, human-in-the-loop controls, deterministic guardrails, and what happens when the model is wrong.

Doing both well is the job.

Why qualification is the other half of the job:

They do not sell licenses. They sell a solution they then have to build, in the client’s environment, at an accuracy level where being 95% right on a $200M position is still wrong.

That means every opportunity you support consumes engineering capacity to scope, and their delivery organization is finite. You are the person who looks at the actual data, the actual workflow, and the actual constraints and says whether this is real. The honest no is one of the most valuable things you will produce.

What you will own:

• Technical discovery. Get underneath the stated requirement to the actual workflow, the exception paths, and where the current process breaks.

• The business-buyer demo. Own the flow, the narrative, and the data, including the Context Studio story. Build it so a Client Partner can run it without you.

• Architecture and security reviews. Lead conversations on data access, deployment model, controls, lineage, and auditability.

• Model risk and compliance. Support the questions that stall deals in regulated institutions.

• Accuracy and evals. Explain, credibly and specifically, how they get from a frontier model’s baseline to production accuracy — ontology, encoded SOPs, edge-case handling, error correction, and the eval sets that prove it. Technical buyers will test this, and hand-waving ends the conversation.

• Scoping with engineering. Translate opportunities into estimates the delivery team can stand behind, and be honest with sellers about what capacity allows. You produce the opportunity brief, the ROI validation, and the scope that feeds the SOW.

• Clean handoff. The delivery pod inherits your scope. Make it something a forward-deployed engineer can build from without re-discovering the account.

• Technical enablement. Build the objection library, the proof points, and the talk tracks so the commercial team is not dependent on you for every call.

What success looks like:

90 days: You run technical discovery and demos without a founder in the room. A working business-buyer demo exists.

6 months: A repeatable discovery framework an…