System Engineering In

Forward Deployed Engineers Hybrid, New York (Manhattan) | Contract | Competitive day rate

We are unable to sponsor visas, and we do not consider visa holders working through a C2C arrangement.

We are looking for a number Forward Deployed Engineers to work alongside our AI Solutions Lead, building production AI applications directly for our clients.
You will partner with the client to understand their operational problems, then design and ship full-stack solutions powered by LLMs and agentic workflows. This is a client-facing role, not a back-office build role. You will own projects from discovery through deployment, working closely with the client team and iterating fast as requirements shift.

Build and deploy production-ready applications with modern full-stack tooling
Develop LLM features, AI agents, and intelligent workflows inside customer environments
Build backend services and frontend experiences that hold up under real use
Run prompt engineering and evaluation cycles to keep improving what's shipped

Solid background in software engineering or AI engineering, ideally with roots in solutions engineering, implementation, or embedded delivery work before AI made it fashionable

• 8+ years in software engineering overall
• 5+ years in full-stack development (React, SQL, and Python or TypeScript as your main language, with working fluency in the other)
• 2+ years hands-on shipping LLM or generative AI applications in production
• 1+ year designing RAG pipelines and working with vector databases
• Genuine interest in where this field is heading, and the ability to adapt fast when the tools change
• You'll be the engineer customers see and trust

Core languages and backend
Python: for AI orchestration, data scripting, and backend logic
TypeScript / Node.js: for customer-facing extensions, full-stack glue code, and lightweight web UIs
React / Next.js: for customer-facing interfaces
SQL: for deep database queries (Java or Go is a plus for heavy enterprise system extensions)
RAG and vector stores: hands-on design with vector databases (Pinecone, Qdrant, pgvector) and retrieval systems
Evaluation and observability: deployment tracking using tools like LangSmith or Braintrust to audit hallucination rates and latency
Data and infrastructure
Databases and warehouses: PostgreSQL, Snowflake, or BigQuery for customer data ingestion
Cloud platforms: working knowledge of AWS, GCP, or Azure
While I carefully read each one, I am unfortunately unable to reply to everyone individually due to the volume of applications.