Descripción del puesto

Information about project

The Forward Deployed Engineer (FDE) is responsible for the technical success of our most critical AI implementations. You will bridge the gap between our core AI infrastructure and the bespoke needs of our largest users. You are part software engineer, part architect, and part field-ops, tasked with ensuring our AI capabilities solve real-world problems at scale.

Requisitos

Responsibilities

Embedded Implementation: Join "tiger teams" to deploy our AI tech stack within specific business verticals or client environments, ensuring 1:1 fit between our API and their workflow. Architectural Guardrails: While moving fast, you will ensure all "last-mile" deployments adhere to our internal standards for security, data privacy (PII masking), and LLM cost-management. Tooling Development: Identify patterns in "one-off" client requests and build internal libraries, SDKs, or scaffolding to automate those tasks for the rest of the engineering team. Technical Discovery: Partner with Product Managers to conduct technical "deep dives" with users, uncovering the friction points that standard telemetry misses.

Must have

The Stack: Professional experience with Python, Node.js, and modern AI orchestration frameworks (e.g., LangChain, Semantic Kernel, or AutoGPT). Infrastructure: Deep understanding of vector databases, embedding models, and the nuances of prompt engineering (few-shot, CoT, etc.). Integration: Ability to navigate complex enterprise auth (OAuth, SAML) and integrate with disparate data silos (SQL, NoSQL, ERPs). Testing & Eval: Experience building evaluation frameworks to measure LLM performance, hallucination rates, and latency. The "Soft" Profile High Agency: You don't wait for a spec. You find the problem, prototype the solution, and get buy-in. Translational Skill: You can explain why a RAG pipeline is failing to a non-technical project lead without losing them in the jargon.

Nice to have

AI Reliability (Evals): Experience using tools like LangSmith or Promptfoo to test, benchmark, and fix AI "hallucinations." Data Wrangling: Ability to extract and clean text from messy real-world files like complex PDFs, PowerPoints, and legacy databases. Advanced Retrieval (RAG): Deep knowledge of making AI "smarter" by connecting it to internal company data via Vector Databases. Agentic Workflows: Experience building AI that doesn't just talk, but actually does things (e.g., calling APIs, updating Jira, or sending emails). Speed-to-Market: A track record of shipping working prototypes in days, not months, using AI-native tools like Cursor or Vercel.

Beneficios

• PTO
• Days off
• Salary un USD
• Long term contract