Forward Deployed AI Engineer

arrow

New York / $85 - $100 hour

INFO

Salary
SALARY:

$85 - $100

Location

LOCATION

New York

Job Type
JOB TYPE

Contract

Forward Deployed AI Engineer

Role Overview

Our client is seeking experienced AI Forward Deployed Engineers to support the development and operationalization of an AI-driven document compliance platform. This system uses multimodal AI models to interpret technical documents-such as permits, plans, and regulatory filings-and assess them against defined rule sets.

In this role, you will contribute directly to model prompting, evaluation methodology, system reliability, and cloud‑native deployment. You'll work closely with internal teams and external partners to guide the platform from advanced prototype to production-ready release.


What You'll Do

AI / Model Development

  • Create, refine, andoptimize prompts formultimodal AI systems performing:
    • document categorization
    • structured information extraction
    • rule-based validation workflows
  • Design structured output formats using JSON-based schemas, function-calling techniques, and typed validation frameworks.
  • Assess multiple model providers to compare costs, accuracy, latency, and overall performance, recommending model upgrades as needed.
  • Build repeatable evaluation pipelines that measure model quality using quantitative metrics
  • Strengthen the evaluation workflow by introducing formalized testing and more detailed performance analysis.
  • Develop benchmark datasets and transform current evaluation practices into a fully in-house process.
  • Lead or support expert labeling and assessment efforts (preferred but not required).

System Engineering

  • Maintain and enhance a Python-based service layer built with modern API frameworks and asynchronous execution patterns.
  • Collaborate with frontend teams using TypeScript and modern React frameworks to support end-to-end workflows.
  • Design and scale pipelines across a cloud-native stack leveraging storage, messaging queues, and distributed databases.
  • Diagnose and resolve issues that span AI inference, backend services, and cloud deployment.
  • Work with both language-vision models and classical computer vision tools.
  • Contribute to a production rollout with multiple municipal or enterprise partners over a 3-4 month timeline.

Required Qualifications

  • Strong command of prompt engineering techniques for multimodal or vision‑language AI systems.
  • Experience building structured evaluation frameworks:
    • dataset generation and curation
    • metric-based scoring (e.g., precision, recall, F1)
    • granular per-sample review
  • Understanding of multimodal AI that combines documents, imagery, and text.
  • Advanced Python skills (including async programming, FastAPI, Pydantic, and related tooling).
  • Experience with queue-based architectures (Azure, AWS, or equivalent).
  • Familiarity with distributed storage solutions and NoSQL databases.
  • Ability to work within a modern TypeScript/React application environment.
  • Strong understanding of performance, cost, and latency optimization in AI-driven systems.

Preferred Qualifications

  • Experience with containerization (Docker or similar), reverse proxies, and multi-service orchestration tools.
  • Familiarity with CI/CD pipelines and automated deployment workflows.
  • Experience building rich, data-intensive user interfaces or workflow tools.
  • Understanding of environment management across multiple deployment stages (dev, staging, production).

Location: Remote (U.S. based)
Duration: 3/4-months initially, with high likelihood of extension
Hours: 40 per week
Pay: $85/hr-$100/hr
Benefits: W2 candidates are eligible for healthcare, dental, vision, and 401k

**WE WILL NOT BE UTILIZING THIRD PARTIES FOR THIS SEARCH. WE WILL ONLY SPEAK WITH CANDIDATES DIRECTLY.**


CONTACT

SIMILAR
JOB RESULTS

4k-Harnham_DA copy
CAN’T FIND THE RIGHT OPPORTUNITY?

STILL
LOOKING?

If you can’t see what you’re looking for right now, send us your CV anyway – we’re always getting fresh new roles through the door.