مهندس ذكاء اصطناعي 🤖🚀 #التعلم_الآلي #تكنولوجيا_المستقبل
Dubai
Joined September 2022
- Tweets1.5K
- Following268
- Followers43
- Likes7.6K
Rami Tamer retweeted
Scaling #AI across the Middle East and North Africa will require more than computing power. Energy and water infrastructure must keep pace.
A new @wef report, produced with @SRMGthink, Powering AI at Scale: Optimizing Hyperscale Data C ..
rt @wef
Rami Tamer retweeted
AI investment workflows need more than a confident chatbot.
openInvest is a self-hosted investment decision engine for builders who want to connect modern AI agents to a reviewable investment workflow.
It helps you document and scrutinize a decision by separating macro, quantitative, and risk analysis before producing a structured verdict and audit trail.
Key features:
• Isolated committee roles – separates macro strategy, quantitative analysis, and risk review to enforce information boundaries
• Cross-challenge process – quant and risk analysts review each other’s first-round reports before the final synthesis
• Auditable decision trail – records the reasoning behind each verdict in Markdown and YAML files tracked by Git
• Backtesting guardrails – includes a long-horizon harness with protections against lookahead bias
• Agent-friendly access – offers a plugin, MCP server, and CLI paths for supported agent environments
It’s open-source (MIT license).
Link in the reply 👇
Rami Tamer retweeted
DeepSeek 4.1-1017 is currently #1 CN and open-source model on the AA intelligence benchmark
let’s see see if any of the CN launches this week can top this
- minimax m3.1 flash launched yesterday
- kimi k3.1 and qwen 4 (coming soon)
who’s your bet on to take the top spot ?
Rami Tamer retweeted
Revolut's Achilles Heel: The Price of Being the North Star
Worth a read as operational excellence ia where Revolut is behind.
As the outage is lasting, one has to wonder what the impact will be on
- customers (retail and business)
- regulators that have licensed Revolut
- investors
Rami Tamer retweeted
Stop rewriting the same subagent brief every time you delegate.
Hermes Teammates is a Hermes Agent plugin for builders who want a named roster of reusable subagents, such as a reviewer or scout.
It helps you delegate repeatable work with clearer ownership by carrying each teammate’s instructions, toolsets, and optional model configuration into its assignments.
Key features:
• Named teammate roster – define specialists in config instead of recreating their roles for every task
• Assignment + status tools – launch a teammate, receive a run ID, and check its progress or result
• Bounded follow-ups – pass a digest of earlier assignments so a teammate can continue related work
• Live steering – send instructions to a running teammate without stopping the run
• Kanban review lane – teammates can claim lane tasks and move successful work to review
It’s open-source (MIT license).
Link in the reply 👇
Rami Tamer retweeted
The U.S., India, and UAE have real opportunities to advance cooperation on energy, trade, technology, and regional connectivity. @SecRubio, @DrSJaishankar, and UAE Foreign Minister @ABZayed in Mumbai for a discussion on turning our close ties into concrete results!
Rami Tamer retweeted
Today, we present the Dubai-it Award in the Government category to the Red Carpet project by the General Directorate of Identity and Foreigners Affairs in Dubai. The project has transformed the airport experience, enabling travellers to complete entry and exit procedures in less than five seconds, making it the fastest service of its kind in the world. It has increased capacity tenfold while achieving a passenger satisfaction rate of over 99%.
This is the philosophy behind Dubai-it: achieving exceptional results that change the way things are done, harnessing the latest technologies to make people’s lives easier and create experiences that exceed their expectations.
Congratulations to the Red Carpet project team.
Rami Tamer retweeted
Does your bank really know you? 🏦🤔
ENjpy the full video here: 👉 ow.ly/guwo50ZZbCy
💎 In this video snippet, we dive into how much data financial institutions actually have on us versus the complete picture of our lives. While most people use upwards of 10 to 23 different financial providers (from insurance to loans and credit cards), primary banks usually see less than 3% of your overall personal data!
With concepts like the Quantified Self and Digital Twins rising, gathering personal data across health, lifestyle, and financial habits allows for smarter optimization in work and daily life. But when it comes to banking, they’re only seeing a fraction of the bigger picture.
🤔 What do you think? Should banks leverage broader personal data to give better services, or is data privacy the top priority? Share your thoughts in the comments! 👇
Rami Tamer retweeted
Everyone knows liquidity matters for tokenised assets.
But few have made meaningful progress.
@centrifuge is the standout. 👇
From instant redemption facilities with @grovedotfinance to integrations with Settle by @redstone_defi, @FissionXYZ, MultiLiquid and Symbiotic, the team is tackling this challenge from multiple angles.
The result: 47% utilisation for $JAAA, with JTRSY arguably among the most liquid tokenised Treasury products on the market.
And none of this is easy to build.👇
Rami Tamer retweeted
AI coding tools are easy to try. A disciplined workflow is harder to build.
AI Dev Tools Zoomcamp is a free, hands-on course repository for developers and technical data professionals learning AI-assisted software development.
It helps you move from a vague product idea to shipped software by pairing coding agents with specifications, review, testing, deployment, and guardrails.
Key features:
• AI-native workflow – turn a product idea into a spec, focused backlog, and independently verified implementation
• Full-stack path – build an app from a frontend prototype and OpenAPI contract through auth, real-time collaboration, persistent storage, and tests
• Deployment foundation – containerize with Docker Compose, move to Postgres, add integration and end-to-end testing, and deploy with CI/CD
• Observability module – instrument services with OpenTelemetry, Prometheus, Loki, Tempo, and Grafana, plus an actionable alert
• Agent extension patterns – create reusable skills and focused subagents, then orchestrate work in isolated Git worktrees
Free public GitHub repo.
Link in the reply 👇
Rami Tamer retweeted
According to the 2025 Stack Overflow Developer Survey, 84% of developers are using or planning to use AI tools. 69% of the ones already using AI agents reported a productivity bump.
I trust the first number more than the second. Seeing the same split where you work?
Rami Tamer retweeted
Researchers used ribosomal RNA sequencing fragments from Ribo-seq data to identify changes in ribosome composition and collisions, revealing new insights into protein synthesis and gene regulation. - @NKI_nl - @oncodeinstitute rna-seqblog.com/ribosome-pro…
Rami Tamer retweeted
Don't miss our next live podcast episode! We're excited to have Louis-François Bouchard join us to discuss Shipping Reliable LLMs and the Rise of AI Engineering. Tune in to learn more about their experience and insights.
Register here: luma.com/nassme8x
Rami Tamer retweeted
📱 Bookmark these 10 GitHub repos that let an AI agent use your phone
1. Open-AutoGLM (26K stars, Apache 2.0)
github.com/zai-org/Open-Auto…
A free phone model plus a framework. You type a task, and it taps through Android apps for you. Most docs are in Chinese.
2. Maestro (16K stars, Apache 2.0)
github.com/mobile-dev-inc/Ma…
A mobile test tool with a built-in MCP server. Your coding agent taps through your app, then saves the steps as a test.
3. Midscene (15K stars, MIT)
github.com/web-infra-dev/mid…
One agent API that looks at the screen and acts on Android, iOS, web and desktop. Built for app testing first.
4. UI-TARS (12K stars, Apache 2.0)
github.com/bytedance/UI-TARS
ByteDance's open model that sees a screen and acts, with a mode for phones. It is a model, so you connect it to a phone yourself.
5. ARTEMIS (11K stars, Apache 2.0)
github.com/google/artemis
From Google's GitHub. Give an order in plain English and it does the work on Android. It plugs into Claude Code and Codex through MCP.
6. Mobilerun (10K stars, MIT)
github.com/droidrun/mobileru…
Formerly Droidrun. An agent framework for Android and iOS that works with any model, from a CLI or Python.
7. Mobile-Agent (9K stars, MIT)
github.com/X-PLUG/MobileAgen…
Alibaba's family of phone agents and the GUI-Owl models behind them. Research code.
8. Mobile MCP (9K stars, Apache 2.0)
github.com/mobile-next/mobil…
An MCP server that lets Claude, Cursor or any agent tap and type on iPhones, Android phones and emulators. It sends usage data unless you turn it off.
9. agent-device (5K stars, MIT)
github.com/callstack/agent-d…
A CLI and MCP server so coding agents can control and check your app on iOS, real iPhones included, and Android.
10. mobile-use (3K stars, Apache 2.0)
github.com/minitap-ai/mobile…
An agent that uses real Android apps from plain-English commands and gives back structured data.
One warning: test these on a spare phone or an emulator. Hidden text on a screen can take over a phone agent, so keep it away from your bank and chat apps.
Rami Tamer retweeted
A shared AI platform could solve a problem that individual AI pilots cannot: consistent accountability across an organization.
McKinsey’s public-sector framework illustrates how different use cases, from fraud screening to permit processing, can operate through a common governance layer.
Identity controls, testing, audit logs and model registries are built into that layer rather than recreated for every application.
The strategic benefit is significant. Agencies can develop new AI services while maintaining common standards for access, oversight and evaluation.
This also changes where responsibility sits. Business units remain accountable for their services, while shared technology and governance capabilities establish the safeguards.
Scaling AI responsibly requires more than approving individual tools. It requires an architecture in which governance scales alongside adoption.
Source: McKinsey & Company, “Rewiring the public sector: How controls can accelerate AI”
@RLDI_Lamy @sulefati7 @ILoveBooks786 @Corix_JC @FrRonconii @COSTESLionelEr @Nicochan33 @faryus88 @ramonvidall @ozsilverfox @jameslhbartlett @dinisguarda @bbailey39 @TysonLester @PatrickCFOD @anton_chuvakin @PrimeStata @OperatingByJohn @Eli_Krumova @Khulood_Almani @Transform_Sec @quepasachico
Rami Tamer retweeted
Running AI on your own devices or internal servers is becoming more realistic
🎯 Local Models Are Improving. That Doesn't Mean They Go Everywhere.
When you run a model locally, proprietary code and internal business data don't leave your infrastructure. That's a meaningful shift for companies that handle sensitive information and want to know exactly where it's being processed.
→ gradientflow.com/the-local-a…
Rami Tamer retweeted
🚨 Want to learn how to build + ship AI and Data Science projects (that businesses actually want in 2026)?
On October 21st, I am hosting a free workshop to help you get started with AI + DS projects in Python.
Register here (500 seats): learn.business-science.io/ai… 
