Selected work
AI for property decisions, supported by product engineering, data systems and public tools.
Professional work
Appraiva
- Role
- Co-Founder & Technical Lead · October 2023–present · Cleveland, Ohio; remote from Dubai
- Problem
- Connect property discovery and enriched data with useful investor and developer workflows.
- Approach
- Lead product architecture, data systems, AI workflows and cloud infrastructure; work across customer discovery and product development.
- Output
- Platform capabilities for property discovery, data enrichment, search and opportunity analysis.
Royal Abraj Group
- Role
- Co-Founder & CEO · 2022–present · Dubai
- Problem
- Align commercial priorities, client experience and operating work across property brokerage and documents-clearing services.
- Approach
- Lead digital strategy and technology, translating business needs into product priorities, processes and data ownership.
- Output
- Development of websites, CRM workflows, a customer portal and internal tools for enquiry capture, follow-up, document handling and case tracking.
Oxfam Novib
- Role
- Data Engineer · September 2022–June 2023 · The Hague; remote
- Problem
- Prepare data for CRM migration and analytics.
- Approach
- Build Azure pipelines using Databricks, SQL, Data Factory, Key Vault and Azure DevOps; integrate SAP and Salesforce data.
- Output
- Multi-source ETL across Bronze, Silver and Gold layers, anonymization and retention logic, and documented development, production and handoff workflows.
Hamrah e Aval (MCI)
- Role
- Machine Learning Engineer · November 2020–April 2022 · Tehran
- Problem
- Support database marketing with customer segmentation.
- Approach
- Use RFM (recency, frequency and monetary) analysis and behavioral data; tune segmentation models and memory use.
- Output
- Customer segmentation models supporting churn-risk analysis and campaign targeting. No campaign uplift is asserted.
Public studies and tools
These descriptions summarize my supported professional background. Public tools and studies are separate artifacts; their limits remain part of the work.
- Rent and ownership scenarios: 393 metros — ACS/HUD cash-budget scenarios; not matched-property investment returns.
- Metro rent burden: ACS 2024 — Descriptive point estimates for 386 US metros, with reproducible source data.
- Rental cash-flow calculator — Enter costs and financing assumptions; inspect the resulting cash budget.
- Dubai purchase-cost calculator — Separate purchase-price and mortgage bases; check the applicable official route.
- Valuation range checker (English) — Compare 2–5 estimates; agreement does not prove accuracy.
GitHub
- Evidence Note — Record a claim, its source and the next check; recording does not verify it.
- AI workflow cost ledger — Offline cost per accepted result, including failed attempts and review.
- Video publication-date linter — Offline checks for HTML JSON-LD and video-sitemap dates.
- Rent-burden source — Reproducible ACS 2024 analysis and source data.
- Rent/ownership source — Rebuild the published scenarios and inspect assumptions.