Learning AI engineering in public
AIabdurrahmanfaheem.com
Building open, deployed agent projects with evaluations, tracing and cost controls, and writing up what works and what breaks.
Open to AI engineering conversations
I build AI agents that work in production, not just in demos.
I'm AbdurRahman, a data scientist and engineer in London. I deploy AI agents on Google Cloud and write about what it really takes to move an agent from prototype to production.
- Agent in production
- 1
- Colleagues supported with AI
- 800+
- Pipeline runtime cut
- 2 wks → 1 hr
- Dashboards built
- 29
What I work on
From prototype to production
01
Agents in production
Deploying AI agents that run inside real companies: service accounts, permissions, monitoring and cost control, not just a demo.
02
Google Cloud & Gemini
Vertex AI, the Gemini API, BigQuery and the plumbing that connects an agent to the systems it needs.
03
Automation that pays off
Python and SQL pipelines that turn weeks of manual work into minutes, measured in hours saved.
Projects
Things I've built
Case study
Production compliance-monitoring agent
How an autonomous agent on Google Cloud reads, checks and reports on compliance signals, and the permission and reliability lessons from running it in a real company. Write-up coming soon.
Vertex AI · Gemini API · Gmail API · GCP IAM · Python
In progress
Open-source agent with evals
A small, useful agent built in the open, with automated evaluations, tracing and a cost dashboard, deployed with a live demo.
Python · Google ADK · FastAPI · Docker
Case study
Vision-based lab automation
BEng dissertation: a physical prototype that uses computer vision to detect and count chemical liquid drops without a human watching.
Python · OpenCV · Hardware
Journey
How I got here
- Now
- 2024 – present
Deployed a compliance-monitoring AI agent
AIRVU (Uswitch, Tempcover, Money.co.uk)Designed and deployed an autonomous agent on Google Cloud using Vertex AI, the Gemini API and the Gmail API to automate a recurring compliance reporting workflow.
- 2024 – present
Internal AI consultant
LeadershipRVU · 800+ employeesOne of four internal consultants championing AI adoption: leading training and designing internal AI productivity frameworks.
- 2024 – present
Data infrastructure & automation
EngineeringRVUPiped 13 datasets into BigQuery, built 29 Tableau dashboards, and cut a quarterly SQL aggregation from two weeks to one hour.
- 2024 – present
Machine learning models
EngineeringRVUK-Means customer segmentation, and a predictive investment model in Python and scikit-learn supporting a £10M+ marketing portfolio.