Mohammad Irshad
AI & Backend Engineer
I design the systems that don't make the screenshot — the APIs, pipelines, and LLM infrastructure that stay up while traffic, cost, and complexity try to take them down.
I'm an AI & Backend Engineer with 3+ years building production-grade platforms end to end — from scalable Node.js/NestJS backends and Next.js frontends to generative-AI microservices.
I've shipped e-commerce and delivery systems with real-time chat and video, payments, logistics and tax-compliant invoicing, and engineered RAG pipelines, LLM assistants and computer-vision systems on top of OpenAI and AWS. I care about clean, secure, maintainable architecture that turns complex requirements into reliable products.
Languages: English · Urdu · Arabic
Education
BS in Software Engineering
University of Engineering & Technology (UET), Mardan
Mardan, Pakistan · 2023
- Awarded a merit-based scholarship for academic performance.
A working toolkit across backend, frontend, data, cloud and AI.
Languages
Backend
Frontend
Databases & ORMs
Cloud & DevOps
Real-time & Integrations
Generative AI
ML / Deep Learning
Computer Vision
Security & Tools
AI & Backend Engineer · Confidential
Jun 2025 — PresentDammam, Saudi Arabia
- Backend & AI engineering on a confidential platform — building and operating production services end to end. Details available under NDA.
AI Engineer · Luminous Rays LLC
Aug 2024 — May 2025Remote, United States
- Built a Retrieval-Augmented Generation (RAG) recommendation engine using OpenAI embeddings and Pinecone for semantic search across 10,000+ listings.
- Developed a multi-modal GPT-4o-mini chatbot with persistent MongoDB history, Whisper voice-to-service transcription, and GPT-4 Vision image moderation.
- Architected a multi-modal dispute-resolution system producing automated verdicts with confidence scoring, exposed via a Dockerized FastAPI microservice.
AI Engineer · IR Solutions
Jul 2023 — Jul 2024Islamabad, Pakistan
- Built a YOLO + OpenCV pipeline to automatically remove car-image backgrounds via segmentation, masking, and contour post-processing.
- Developed a hospital assistant chatbot with RAG over a medical knowledge base and OpenAI Function Calling for appointment scheduling.
- Deployed REST APIs over Stable Diffusion for parameterized text-to-image generation and a photo-mosaic image application.
- Built an interactive financial-metrics dashboard (Streamlit/Python) computing key metrics and visualizing trends with Matplotlib/Seaborn.
- Engineered a generative-AI financial chatbot with web retrieval, prompt engineering, similarity search, and multi-turn memory.
- Developed a face-similarity retrieval system using OpenCV and FaceNet/ResNet embeddings indexed in Pinecone for fast top-N matching.
Drag the timeline — or focus it and use the arrow keys — to rewind from today back to 2023. Watch the stack, the role, and the picture itself age.
AI & Backend Engineer · Confidential
Jun 2025 — NowDammam, Saudi Arabia
Backend & AI engineering on a confidential platform, end to end.
- Building and operating production backend and AI systems end to end. Details available under NDA.
Each card is the architecture itself — hover to watch a request flow through it. Most repositories are private; reach out for a walkthrough.
AI Microservices Suite
RAG, multi-modal chat, voice & vision
Treated each model as an isolated FastAPI service with its own memory and confidence scoring — so a bad verdict is observable and contained, not silent.
A Dockerized FastAPI suite powering a marketplace: a RAG recommendation engine (OpenAI embeddings + Pinecone) over 10k+ listings, a GPT-4o-mini assistant with MongoDB memory, Whisper voice-to-service, GPT-4 Vision moderation, and a multi-modal dispute-resolution engine with confidence-scored verdicts.
Applied AI / ML Projects
Vision, RAG chatbots & generative AI
A run of focused proofs — each one taking a single AI capability (retrieval, detection, generation) from notebook to a callable API.
A collection of earlier applied-AI builds: a hospital RAG chatbot with function-calling appointments, YOLO + OpenCV car background removal, Stable Diffusion text-to-image APIs, a generative-AI financial assistant with web retrieval, and a FaceNet + Pinecone face-similarity retrieval system.
Have a role, a project, or an idea? Send a message and I'll reply as soon as I can.
I'm open to backend, AI, and full-stack opportunities. The fastest way to reach me is the form — or any of these:
Download CV