Artificial Intelligence Engineer

Staffworx; New York (NY); Hybrid; USD 175,000 - 300,000; Full time

Job summary

Staffworx is hiring an AI Engineer for the Agentic Platform in New York City. This platform role focuses on building and hardening the agentic layer across products, enabling sustained growth for a venture-backed health tech analytics company.

The role requires 5–10 years in AI/ML, deep experience with evals and harness engineering, and strong machine learning fundamentals. Hybrid work four days onsite in NYC with collaboration across a small, senior team.

Qualifications

• 5–10 years in AI/ML engineering with agentic systems shipped and running in production.
• Deep expertise in evals, harness engineering, context engineering, and tool surfaces.
• Strong foundations in machine learning; explain model architecture and training from first principles.
• Experience delivering end-to-end projects in a startup environment with clear ownership.
• BS or MS in Computer Science, Mathematics, Machine Learning, Statistics or a related quantitative discipline.
• Relational data modelling; experience with GCP Vertex AI and Claude SDK; excellent written and spoken English.

Responsibilities

• Define MCP-style contracts for tools, connectors and skills to add capabilities without destabilising the platform.
• Build connective tissue so agents, tools, context and data compose cleanly across products.
• Own production-grade eval harnesses, replay and benchmarks, and gate releases on them.
• Lead context and harness engineering, grounding agents in customer business rules and data.
• Fine-tune in-house and open-source models, benchmark against frontier baselines, and decide build vs. buy.

Skills

AI/ML engineering

Python

Eval harness engineering

Context engineering

Tool surfaces

Production engineering

Education

BS or MS in CS/Math/ML/Statistics

Tools

GCP Vertex AI

Claude SDK

Job description

Location: New York City, NY, USA. Hybrid, four days on-site (Monday to Thursday)

Job type: Permanent, full time

Salary: $175,000 to $300,000 base, plus performance bonus and competitive equity

Start: ASAP

Sponsorship: Open to visa transfers (OPT, H1B)

Reference: SWX-889456

AI Engineer, Agentic Platform. New York City. $175,000 to $300,000 plus bonus and equity.

Staffworx is retained on an exclusive search for a Series A venture-backed AI company building the financial operating system for US healthcare. The business is around 30 people, headquartered in Manhattan

US healthcare providers account for roughly $2.5 trillion of medical expenditure on margins of two to five per cent, and a meaningful number of provider organisations are at genuine risk of failure. Finance teams are stuck in spreadsheets and disconnected legacy systems. Our client automates that manual data work with agentic AI, so finance leaders can act on current information rather than last month's.

This is a platform role, not a customer configuration role. You will build and harden the agentic layer that every product team builds on top of.

What you will be doing

• Defining the MCP-style tool, connector and skill contracts that let new capabilities be added continuously without destabilising the platform
• Building the connective tissue so agents, tools, context and data compose cleanly across products
• Owning production-grade eval harnesses, replay and benchmarks, and gating releases on them
• Leading context and harness engineering, grounding agents in each customer's business rules and data
• Fine-tuning in-house and open-source models, benchmarking against frontier baselines and making the build versus buy calls

What we need to see

• Five to ten years in AI/ML engineering, primarily Python, with agentic systems shipped and running in production
• Real depth in evals, harness engineering, context engineering and tool surfaces, rather than framework-level familiarity
• Rigorous machine learning foundations. You should be able to explain model architecture and training from first principles at a whiteboard
• Experience building end to end in a startup environment, with clear evidence of independent ownership
• BS or MS in Computer Science, Mathematics, Machine Learning, Statistics or a closely related quantitative discipline
• Relational data modelling. GCP and Vertex AI are a strong plus, as is experience with the Claude SDK
• Excellent written and spoken English. The culture is documentation driven and opinions are solicited in writing

Culture and working pattern

Small, deliberately senior engineering team with a flat, democratic structure and RFC-style decision making. Heavily agentic development process with strong human judgement layered on top. Four days a week in the New York office, Monday to Thursday, with Fridays worked from home. The team is hiring two to three people into this role.

Selection process

Twenty minute introductory call with the Head of Engineering, a two hour take-home machine learning exercise, a one hour remote review of that exercise with two AI engineers, then a single on-site final round in New York covering technical and behavioural interviews, a product conversation and CEO sign-off. Offers follow quickly.