Forward Deployed Engineer, AI

ClarityPay•New York, NY

2d•$145,000 - $195,000•Hybrid

About The Position

ClarityPay is seeking Forward Deployed Engineers to embed with the teams that run the company and turn how they work into dependable, AI-native systems. You will sit with process owners in finance, credit and portfolio risk, servicing and collections, and operations; scope the problem with them before building; ship agents, controls and automations into production; and stay through adoption as the named secondary owner of the process. You will join the Data & AI organization, working alongside engineers in New York and India on a shared platform, and alongside Product and Engineering you will build external-facing AI solutions for merchants and partners on the ClarityPay platform. You will use AI tooling to make the build the easy part, and you will be measured on what is adopted and relied upon, not on what is demoed.

Requirements

Curious, energetic, positive, solution-oriented, collaborative, trusted.
2–5 years of software engineering or technical deployment experience that includes working directly with business users — in operations, business systems, solutions engineering, or technical consulting.
Production experience with LLMs: prompt engineering, agent or skill development, and evaluation. Things other people depend on, not demos.
Strong Python and SQL: you write and review production code and use AI-assisted development to move faster. Low-code and automation platforms (e.g., Retool, Zapier, n8n) are useful, not a substitute.
Data fluency in a lakehouse environment (e.g., Databricks, ClickHouse): pull answers out of governed data, build reliable pipelines, and catch when the output is wrong.
High agency and comfort with ambiguity — you will choose what to automate as often as you are told.
Strong discovery and communication skills: able to sit with a collections specialist or a finance analyst, understand their workflow, and explain trade-offs to a non-technical audience; comfortable working across time zones with engineering colleagues in India.
Judgment about what should not be automated — where the cost of being wrong outweighs the time saved.

Nice To Haves

Fintech, lending, payments, banking, or another regulated environment.
Experience building MCP servers, sub-agents, or agent skills.
MLOps or experimentation experience: evaluation frameworks, staged A/B tests, model monitoring.
Prior forward-deployed, professional services, or founding-engineer experience.

Responsibilities

Embed with process owners to map how the work actually runs, scope the problem with them before building, and agree success criteria and a baseline up front.
Own delivery end to end — discovery, technical scoping, system design, build and production rollout — and stay through adoption as the named secondary owner of the process.
Choose the simplest dependable solution: simplify or remove a step before automating it, then pick a scheduled job, a triggered job, or a bounded agent with a defined mandate, action set, escalation owner and pause conditions.
Write evaluations for everything you ship: prove the output is correct before a team depends on it, and re-check it as models, data and processes change.
Instrument what you build, and the processes around it, with detective (“day-zero”) controls — thresholds, exception inventories and alerting, each with an action path and a named owner — so a breach surfaces the same day, not at month-end.
Build external-facing AI solutions for merchants and partners with Product and Engineering; anything external ships through Engineering’s review and deployment standards.
Codify reusable patterns, evals and harness standards so the second deployment is faster than the first; prefer reusable capabilities over one agent per task, and bring field signal back to the platform roadmap.
Work inside ClarityPay’s controls for AI — evaluate before exposure, bound the exposure, keep a person at the gate — and propose where a human must review, where automation may act unattended, and what gets logged.
Own the value case: set impact hypotheses and baselines, measure before and after, and report to executives on business impact, total effort saved (including review and maintenance), control quality and reuse.
Continually promote and apply ClarityPay values; be a role model of trust and integrity.