Ahmed Mohamed Eid
AI Solutions Architect · Riyadh

From AI ambition to measurable production value.

Most organizations know they need AI, but few know where to start or how to make it pay off. I help them get from ambition to measurable production value — blending deep technical expertise with 18+ years of enterprise judgment.

  • 18+ years in enterprise technology
  • M.Sc. in AI, UT Austin
  • AWS & Google Cloud professional certifications
  • Palantir Foundry & AIP certified

How I take AI from idea to production

01
Identify use cases
Find the places AI is most profitable to apply.
02
Plan the infrastructure
Architect the platform it will run on.
03
Build robust pipelines
Engineer data and retrieval pipelines that hold up in production.
04
Deploy with security & monitoring
Ship with strict access control, lineage, and telemetry.

Selected work

Palantir Foundry & AIP

Enterprise GenAI System — PIF

Architected and built an end-to-end enterprise GenAI system for the Public Investment Fund (PIF) AI solutions team on Palantir Foundry & AIP: RAG across vector, hybrid, and graph-based retrieval; ASR ingestion and transcription; agentic workflows; RBAC, lineage, and telemetry; plus LLM evaluation and regression-testing suites covering retrieval relevance, groundedness, latency, and cost. Led and mentored a team of data engineers.

TONOMUS / NEOM

NEOM Enterprise Architecture

Enterprise AI and cloud reference architectures, landing zones, HLD/LLD, NFR and cost/option analyses, procurement and technology advisory, and AI discovery workshops across NEOM sectors — Food, Environment, Energy & Water, Hotels, and University of NEOM.

UT Austin

Medical Imaging Diagnostic Pipeline

Brain-cancer classification pipeline with Grad-CAM and SHAP explainability. Structured error analysis reduced false positives by 13%.

UT Austin

CNN vs. Vision-Language Models for Explainable Medical Imaging

Compared a CNN's heatmap explanations with LLaVA-Med on brain-tumor MRI. With 4-shot prompts, the vision-language model correctly classified 9 of the 13 scans the CNN got wrong, and explained why.

UT Austin

A GPT Built and Trained from Scratch

A 30M-parameter decoder-only Transformer written from scratch in PyTorch and trained on 50,000 children's stories, with KV caching, mixed precision, and a cosine learning-rate schedule.

View on GitHub
UT Austin

Predicting Apartment Rental Prices in Riyadh

A paper and pipeline predicting annual rents from 8,252 listings, using LLMs to extract amenity features from Arabic listing descriptions. Best model: R² 0.735 on the test set.

Side builds

kumo

AWS-like mini cloud a developer can start in 15 minutes.

ibex

iPhone Backup Explorer.

AI Tender Evaluation

Document analysis and decision support for evaluating technical proposals submitted to an RFP.

Connected Vehicles In progress

Raspberry Pi + GSM/GPS + OBD-II telemetry to an IoT platform.

Latest writing

About

I'm an AI Solutions Architect with 18+ years across telecom and enterprise technology in four countries — Egypt, Morocco, Iraq, and Saudi Arabia — now focused on generative AI and agentic systems for enterprise delivery. I hold an M.Sc. in Artificial Intelligence from the University of Texas at Austin (GPA 3.93), completed May 2026.

Software Engineer — Xpress Integration & Ericsson

Cairo, Egypt · 2008
Telecom BSS/OSS systems.

Consultant — Atos

Rabat, Morocco · 2012
Maroc Telecom.

Senior Consultant — Ericsson

Sulaymaniyah, Iraq · 2013
BSCS iX billing transformation for Asiacell.

Solution Architect — Mobily

Riyadh, Saudi Arabia · 2017

AI & Cloud Solutions Architect — TONOMUS (NEOM)

Riyadh, Saudi Arabia · 2021–present
Embedded with the Public Investment Fund (PIF) AI solutions team.

Ready to move past AI exploration?

Let's talk about where AI can pay off in your organization.

© 2026 Ahmed Mohamed Eid · Built with Astro, hosted on Cloudflare
Theme based on Astrofy (MIT)