MSMuhammed Salim

AI Engineer · Forward Deployed Engineer

PRODUCTION AI
BEYOND DEMOS

 

AI Engineer building production AI systems end to end — multi-agent assistants, RAG, MCP tool layers, and the full-stack + data platforms underneath.

AI Engineer @ CCGAWS · GCP · Azure AI certified

AI Systems Console

connected
assistant

ask me anything about Muhammed — his work, Nexhost, his stack — or general AI-engineering questions (agents, RAG, MCP, production architecture). I'll answer in a systems-minded way.

examples:

promptask about Muhammed, AI engineering, or your system.
$

Experience

Mar 2025 – Present

InsureTech

current

AI Engineer

Claims Consortium Group (CCG)

  • Led product discovery and delivery of a production AI forecasting and scenario-planning agent for senior executives — real-time P&L replacing spreadsheet-driven finance reporting across multi-million-dollar portfolios.
  • Owned end-to-end system and agent architecture with Azure AI Foundry, LangGraph, and Python/SQL — orchestrator and tools for budget projections, cost variance, and revenue forecasting.
  • Delivered the full stack (UI/UX, APIs, data flows), owned budget/ongoing costs, and took the service to production on Azure with GitHub Actions CI/CD.
  • Standardised schemas and automation covering ~70% of the monthly MI cycle — reporting from days to hours, ~40% fewer data errors.
Next role

Jan 2025 – May 2026

AI Engineer (Part-time)

Nexhost AI

  • Sole architect of the NEX platform (TypeScript, React, Python, RAG, MCP, PostgreSQL) on Vercel — zero to live production within weeks.
  • Designed a multi-agent orchestration system from inventory counting to staff rota — ~30% better inventory accuracy, ~25% less food waste, ~70% less scheduling time.
  • Led a 3-engineer team through sprint planning, code reviews, and technical decisions; onboarded beta clients through a successful initial launch.
Next role

Nov 2021 – Dec 2024

Software Engineer

Perkk

  • Developed and maintained the core web platform with JavaScript/TypeScript, React, and Node.js — reusable UI components, APIs, and client integrations.
  • Owned features end-to-end across frontend, backend, and integrations used daily by the delivery team and clients.
  • Helped scale the engineering/delivery team by 8 while keeping quality high; contributed to partner revenue growth and funding wins post-engagement.

Projects

all repos

The stack

AI orchestration

Multi-agent systems with LangChain / LangGraph, tool-calling via a Model Context Protocol (MCP) layer, and RAG-grounded answers — in production, not demos. Resilient multi-provider routing with automatic failover.

Production systems

Python, FastAPI, React / Next.js, TypeScript. Supabase / Postgres with Row-Level Security, multi-tenant design, POS integrations. Config-driven, env for secrets, structured logging, health checks.

Observability

Per-call usage, latency, and cost metering across every model call. Provider health tracking and failover so AI features stay available and economically viable at scale.

Computer vision

Photo-based inventory counting and waste detection with confidence scoring and cost-impact estimation — turning a phone camera into a data-capture pipeline for operations.