CubePeaks | Karachi, Pakistan | Remote / Hybrid (with potential UAE relocation) | Full-time
About CubePeaks
CubePeaks is a UAE-based Microsoft Solutions Partner delivering AI-powered cloud solutions for manufacturing, production, and supply chain businesses across the Middle East. Beyond Dynamics 365 and Power Platform implementations, we build production AI systems that run on real factory floors. These include diagnostics that help technicians resolve machine issues offline, automated supplier quality audits, AI-driven shift handover logging, and intelligent warehouse inspections that extract label data and generate evidence-based reports.
We're growing our AI engineering team in Karachi and looking for a senior engineer who ships AI to production, not just to demos.
The Role
As a Senior AI Full Stack Engineer, you'll design, build, and deploy agentic AI systems end to end, from backend services and agent orchestration to user-facing interfaces and cloud infrastructure. You'll work directly on client solutions for industrial operations where reliability, data privacy, and real-world usability matter.
What You'll Do
- Architect and build agentic AI applications using LLMs, tool calling, multi-agent workflows, and agent harnesses
- Build and integrate MCP (Model Context Protocol) servers and clients to connect AI agents with enterprise systems such as Dynamics 365, ERPs, databases, and internal tools
- Develop robust, scalable backend APIs in Python with FastAPI
- Build RAG pipelines, including document ingestion, chunking, embeddings, vector search, and retrieval evaluation
- Design AI-driven automations that replace repetitive manual workflows for manufacturing and supply chain clients
- Build clean, responsive front-ends for AI products (React / Next.js or similar)
- Deploy, host, and monitor AI solutions on Azure and/or AWS with a focus on security, cost, and performance
- Implement observability, evaluation, and guardrails for LLM applications in production
- Collaborate with consultants and client stakeholders to turn operational problems into working AI solutions
- Mentor mid-level engineers and contribute to engineering standards and best practices
What We're Looking For
- 5+ years of software engineering experience, with at least 2 years building LLM/GenAI applications
- Proven track record of AI solutions built, deployed, and running in production (please share examples)
- Strong Python skills and deep hands-on experience with FastAPI
- Hands-on experience with agentic AI frameworks such as LangChain, LangGraph, LlamaIndex, CrewAI, Semantic Kernel, or AutoGen
- Practical experience with MCP and agent tool-integration patterns
- Experience with agent harnesses and SDKs (e.g., Claude Agent SDK, OpenAI Agents SDK)
- Experience with workflow automation tools and patterns (n8n, Power Automate, Azure Logic Apps, or custom pipelines)
- Solid familiarity with Azure (Azure OpenAI / AI Foundry, App Service, Functions, AKS, Cosmos DB, AI Search) and/or AWS (Bedrock, Lambda, ECS/EKS, S3)
- Experience with vector databases (Azure AI Search, Pinecone, pgvector, Qdrant, etc.)
- Front-end proficiency in React, Next.js, or a similar modern framework
- Comfort with Docker, CI/CD pipelines, and Git-based workflows
- Strong understanding of API design, authentication, and data security
Nice to Have
- Experience with Microsoft Dynamics 365, Power Platform, or Dataverse integrations
- Background in manufacturing, supply chain, or industrial domains
- Experience with LLM evaluation and observability tools (LangSmith, Langfuse, Arize, etc.)
- Exposure to computer vision, OCR, or speech-to-text solutions
- Experience with on-device or offline-capable AI deployments
- Microsoft Azure AI certifications (AI-102, AZ-204)
Why Join CubePeaks
- Work on real AI products solving real problems in live industrial environments
- Opportunity to relocate on-site to our UAE office down the line, based on performance and business needs
- Remote/hybrid flexibility from Karachi with an international, UAE-based company
- Exposure to enterprise clients across the Middle East
- Competitive compensation aligned with experience
- A fast-moving team where your technical decisions shape the product