Contract

Zohorecruit; Piscataway Township (NJ); On-site; USD 140,000 - 190,000; Full time

Job summary

Zohorecruit in Piscataway, NJ is seeking a senior AI software engineer to design, build and harden agentic AI applications that read from and write to enterprise systems of record, grounded on Microsoft Fabric and Azure AI services.

You will lead a small, cross-functional pod from problem framing to prototype and production, running rapid 1–2 week iteration loops, and owning end-to-end quality, security, reliability and governance.

Qualifications

• Proven experience building agentic or multi-agent AI systems in production.
• Strong command of state management, concurrency, idempotency, and authorization.
• Hands-on integration with enterprise APIs and data platforms.
• Experience with modern AI patterns and end-to-end cloud ownership.
• Ability to prototype rapidly and iterate under tight timelines.

Responsibilities

• Design, build and harden agentic AI applications using Fabric and Azure AI services.
• Embed as technical lead of a small pod from problem to production with fast iterations.
• Run iterative cycles with weekly loops and ensure end-to-end quality, security and observability.
• Choose low-code or custom code where appropriate, ensuring safe write-backs to source systems.
• Ensure AI data is governed, lineage documented, and access-controlled for audits.

Skills

Agentic AI apps

Azure cloud

Rapid prototyping

Multi-agent orchestration

Tools

APIs

Data platforms

MCP

Job description

Piscataway, United States | Posted on 09/03/2026

We are a technology services company focused on data, analytics andintelligent robotic processes.

Job Description

WhatYou Will Do

• Design, build and harden agentic AI applicationsthat read from and write to enterprise systems of record, grounded onMicrosoft Fabric and Azure AI services.
• Embed with a Corporate Function team as the technicallead of a small, multi-functional pod — moving from problem, to prototype,to production at pace.
• Run a tight, iterative build cycle: align on theoutcome, prototype, gather feedback, refine and finalise, typically withinone to two weeks per loop.
• Own the quality of what you ship end to end — security,reliability, scalability, observability, and safe, governed write-back tosource systems.
• Make the right build call for the problem: low-code andagent-builder platforms where they fit, custom code where they do not.
• Ensure the data behind each solution is AI-ready —modelled, governed and documented in Fabric, with lineage and accesscontrol that will stand up to audit.
• Set and document the reusable technical patterns,components and guardrails the team applies across every engagement.
• Become credible in a new business domain quickly, andchallenge requirements constructively to land on the right solution ratherthan the requested one.

Requirements

RequiredQualifications

• Proven experience building agentic or multi-agent AIsystems that run in production and serve real users — working systems, notdemonstrations.
• Strong command of state management, concurrency,idempotency and authorisation for systems that act on a system of recordon behalf of many users.
• Hands-on integration experience with enterprise APIsand data platforms, including designing safe read and write paths andhandling failure gracefully.
• Working knowledge of modern AI application patterns —retrieval-augmented generation, prompt and context design, agentorchestration, tool use and protocols such as MCP.
• Strong, hands-on command of a major cloud platform,ideally Azure — able to independently provision, configure, deploy, secureand operate cloud resources, owning an application end to end rather thanonly consuming pre-built services.
• Demonstrated ability to prototype rapidly, usingAI-assisted development and fast iteration to turn a loosely defined ideainto something usable in days.
• The judgement and autonomy of a senior individualcontributor — able to make sound architecture and feasibility calls alone,under time pressure, and explain them to non-technical stakeholders.