AI Forward Deployed Engineer
A forward deployed engineering role at a Series A AI fintech whose product is already live with financial services clients. You sit directly with hedge funds and investment managers, scope the problem on the ground, then build the agentic workflow yourself. Roughly 80% building, the rest managing demanding front-office stakeholders. Early enough that your work shapes how deployments are delivered and what good looks like for enterprise agentic AI in finance.
The work
Most AI roles sit too far from the problem. You join a product team, wait for feedback to trickle through, then ship something weeks later. This is different. You sit directly with hedge funds and investment managers, listen to what portfolio managers, traders and operations teams are trying to solve, then build the workflow yourself. One day that might mean connecting internal data sources. Another, building an agentic workflow that turns fragmented market, client, CRM, email and risk data into something a front-office team can actually use. You won't hand requirements back to a product team. You're the technical lead on the ground: you scope the problem, build the solution, manage the client, handle the pushback and ship production workflows in short feedback loops.
What you'll own
This suits a real software engineer who likes being close to the user. You'll spend most of your time building, but you'll also need the judgement to manage demanding stakeholders when the pressure is high and the answers aren't clean.
- Scoping problems directly with portfolio managers, traders and operations teams, then building the workflow end to end
- Shipping production agentic workflows that turn fragmented market, client, CRM, email and risk data into usable front-office tooling
- Owning the deployment on the ground: build, client management, pushback and short feedback loops
- Shaping how deployments are delivered and what good looks like for enterprise agentic AI in finance
What you'll need
You've shipped production agentic AI systems, not demos or internal experiments, and you can talk clearly about eval suites, failure modes, retrieval issues, tool use and multi-step reasoning, including what you changed when the system didn't behave. You've probably come from forward deployed engineering at a data, AI or enterprise software company; financial services technology across banking, hedge funds, asset management, risk, reference data or market data; or an AI-for-finance startup where you owned client-facing builds. This won't suit a pre-sales engineer, solutions architect, AI strategist or data scientist who hands work to engineering. Roughly 80% of the job is building.
- Production agentic AI systems shipped, not demos or proofs of concept
- Strong Python with hands-on LangChain or a similar agentic framework
- Worked directly with clients or internal business users in high-pressure environments
- Broader stack: RAG, vector databases, LLMs, API design, CRM integrations, SQL, evaluation suites and financial data workflows
- Based in New York City or able to relocate. The four-days-on-site requirement is non-negotiable
What happens next
CV review, a screening call, a technical discussion, a client and team interview, then a final interview before offer. We start with a confidential conversation and share the full picture, including the company, product, hiring manager context and what the first deployment is likely to look like, before anything moves forward.
If it is a fit, you will get a straight read on the process and the comp. If it is not, I will tell you that too.