Mohamed Homaid

I built a verifiable agentic SOC. Here's what I learned about AI-native products. Not "AI-powered." Not "AI-enhanced." AI-native — meaning the agent IS the product, not a feature bolted onto an alert queue. 8 phases. 12 Rust crates. 1 Next.js workbench. What's in it: → Rust streaming substrate (collector → Redpanda → Arroyo + RisingWave → Iceberg) → Hybrid Triager: XGBoost classifier (sub-30 ms) + LLM investigator with MCP-governed tools → Ed25519-signed attestation envelopes — every agent decision is cryptographically reproducible → Live reasoning stepper in a Bloomberg-density workbench The thesis I want to test publicly: most "AI-native" enterprise products today are LLMs grafted onto schemas designed for humans clicking buttons. The real opportunity is rebuilding the substrate so the agent is a first-class operator — governed, bounded, attested. The next 12 months of enterprise software will be defined by who rebuilds vs. who retrofits. Architecture diagram + 5-page brief in comments. Curious — is your industry seeing AI-native rebuilds yet, or still in the wrapper era? #AIAgents #RustLang #CyberSecurity #MachineLearning #SoftwareEngineering #AIInfrastructure #SIEM #OpenSource #TechLeadership #EngineeringLeadership

Dear America, Our tax dollars, military, intelligence, and majority of elected leaders are enabling this. THIS. Are you too racist, ignorant, or cold hearted to demand THIS to stop?

@AmmarCodes Agreed. If the person building it knows what they are doing they could easily add it to ui.shadcn.com/docs/rtl

معظم تطبيقات الويب المولدة بالذكاء الصناعي تستخدم مكتبة Tailwind، لكن هذه التطبيقات المولدة أو حتى اليدوية عادة لا تستخدم طرق التصميم التي تدعم اتجاهين RTL/LTR. لا تظهر هذه المشاكل إلا عند التحويل بين اللغات (عادة في مراحل متقدمة من التطوير). يمكن لهذه الإضافة على eslint من اكتشاف هذه المشاكل مبكراً وحلها تلقائياً. ثلاثة أسطر، "اشبك وانطلق"!

@FahadCoding @rkiza_sa @usesndr موفقين ان شاء الله

أخيراً 🇸🇦 نطلق لكم أول مزود إيميل للمبرمجين من السعودية، وبخطة مجانية تبدأ بـ 3000 إيميل مجاني بالشهر. نستخدمها في كل منصات @rkiza_sa حالياً. فعلا المشروع استنفذ وقتنا وجهدنا بشكل كبيييير, ولكن الحمدلله طلعنا بنتيجة في وجهة نظري اشوف انها رهيبة. @usesndr sndr.sh

This resonates with what I have been seeing building agentic systems across security, FinTech, and healthcare. The shift from selling products to delivering outcomes is the same shift inside the technical architecture. Most agentic pilots work because the context is hand-curated by the engineer in the room. Production breaks because no one ever wrote down what the agent actually needs to receive to do the work reliably. The teams that operationalize AI at scale will be the ones who treat the agent's data contract as the primary deliverable, not the model itself. Forward-deployed engineering is exactly the shape that work takes. Excited to see HUMAIN moving in this direction.

Since May last year, we have been building game-changing products and platforms at @HUMAIN with one core belief: in the age of AI, success is no longer about product sales alone. It is about value realization. Enterprises today are not looking for more software. They are looking for partners who can help them solve real business problems, transform workflows, reduce complexity, and unlock measurable outcomes. This is why our collaboration with @Accenture is so important. Together, we are strengthening HUMAIN’s service delivery capabilities around our core AI offerings through a forward-deployed engineering model where teams work side by side with customers to co-develop solutions that directly address real operational challenges. The future of AI will be won by those who can operationalize AI at scale and deliver tangible business impact. This is the next phase of enterprise AI.   From selling products to delivering outcomes. Read More: lnkd.in/gUKWJw7V #HUMAIN #TheEndOfLimits

Check out my latest article: Why Memory Is the Hard Part of Building Agents linkedin.com/pulse/why-memo… via @LinkedIn

@araghougassian Hey Ara

@m__homaid your DMs are closed!

we’re hosting a 14 day founder program start from nothing build a working product make your first dollar online comment “BET” if you wanna join

We post-trained a 3B model with RL to beat Opus on spreadsheet retrieval. Faster, cheaper, more accurate. - If a piece of your agent loop is narrow, verifiable, and highly repeatable, a tiny trained model might beat the frontier. - The application layer is still early and new verticals are opening fast. Cheap domain specialists orchestrated by a frontier model that only spends tokens on judgment is a bet worth watching.

Daily user. The gap between Claude Code, Codex, and Cursor isn't accuracy, it's whether the model knows when to stop and ask. Claude Code asks. Codex charges ahead. Cursor sits in between. That's where engineering taste actually shows up. Applied. Spent last week building attest-bench.up.railway.app, agentic plane with signed verdicts

I'm looking for 50 more software engineers to join my team at g2i.ai for the next week. We work with frontier labs to help them train models. 100 - 200 USD /hr. Fully remote. Hiring in 150+ countries. RT's appreciated! jobs.ashbyhq.com/g2i/c07a8f96-d…

Full tech stack. No JVM. No monorepo bloat. People keep asking what I used. Here's the whole thing, by layer: **Infra** Docker Compose · Redpanda (Kafka) · MinIO (S3-compatible) · Postgres 16 · Alembic migrations **Streaming runtime (all Rust, no JVM)** Arroyo — ETL + CEP · RisingWave — hot materialized views · Apache Iceberg — warm tier **Backend (Rust workspace, 12 crates)** Axum · Tokio · rdkafka · sqlx · Serde · OpenTelemetry · Anyhow / Thiserror **AI / Agentic** Custom MCP gateway · Anthropic + OpenAI via inference router · Policy engine · Calibration & shadow-check harness · Ed25519 attestation envelopes **ML** XGBoost via ONNX Runtime · SHAP attributions · Novelty detection **Frontend (Next.js 16 App Router)** React 19 · TypeScript · Tailwind · shadcn/ui + REUI · Zustand · @tanstack/react-table · Better Auth + Kysely **Observability** OpenTelemetry → Tempo · Sentry (client + server + edge) · PostHog · Lighthouse CI · Playwright Why this combo: speed, no JVM tax, AI-native by construction, every tool earns its slot. I'll do a deep-dive thread on each layer over the next two weeks. Anyone running a similar stack? Curious what you'd swap. #RustLang #TechStack #NextJS #StreamingData #AIInfrastructure #MLOps #OpenTelemetry #Kafka #SoftwareArchitecture #WebDevelopment

I built a verifiable agentic SOC. Here's what I learned about AI-native products. Not "AI-powered." Not "AI-enhanced." AI-native — meaning the agent IS the product, not a feature bolted onto an alert queue. 8 phases. 12 Rust crates. 1 Next.js workbench. What's in it: → Rust streaming substrate (collector → Redpanda → Arroyo + RisingWave → Iceberg) → Hybrid Triager: XGBoost classifier (sub-30 ms) + LLM investigator with MCP-governed tools → Ed25519-signed attestation envelopes — every agent decision is cryptographically reproducible → Live reasoning stepper in a Bloomberg-density workbench The thesis I want to test publicly: most "AI-native" enterprise products today are LLMs grafted onto schemas designed for humans clicking buttons. The real opportunity is rebuilding the substrate so the agent is a first-class operator — governed, bounded, attested. The next 12 months of enterprise software will be defined by who rebuilds vs. who retrofits. Architecture diagram + 5-page brief in comments. Curious — is your industry seeing AI-native rebuilds yet, or still in the wrapper era? #AIAgents #RustLang #CyberSecurity #MachineLearning #SoftwareEngineering #AIInfrastructure #SIEM #OpenSource #TechLeadership #EngineeringLeadership

25 signs your VIBE CODED app will BREAK at 500 users : Save this before you go live ! 1/ no load testing before launch > you don't know where it breaks because you've never pushed it > one traffic spike and you're debugging live in production 2/ session data stored in server memory > works on one instance > breaks the moment you need two 3/ file uploads going directly to your app server > disk fills up. server dies. files lost. > move uploads to object storage on day one 4/ synchronous email sending in API routes > slow email provider = slow API response for every request that triggers one > offload to a queue. always. 5/ no queue system for background tasks > everything blocking > one slow task pauses everything behind it 6/ hardcoded secrets in deployment scripts > sitting in your CI logs > visible to anyone with pipeline access 7/ single database with no read replica > all reads and writes hitting one machine > first real traffic spike kills query performance 8/ no CDN in front of static assets > every image served by your app server > 500 concurrent users = 500 image requests hitting your backend 9/ DB migrations running automatically on app start > two instances deploy at the same time > both run migrations. race condition. data inconsistency. 10/ no database backup ever tested with a restore > you have backups > you've never actually restored from one 11/ unindexed foreign key columns > every JOIN is a full scan > slow at 100 rows. broken at 100,000. 12/ no rate limiting anywhere > 500 users. one of them is a bot. > your server is now a bot server 13/ API responses with no compression > JSON payloads sent uncompressed > 10x the bandwidth they need to use 14/ no error alerting configured > app crashes at 3 AM > you find out when a user emails you at 9 AM 15/ transactions not used for multi-step writes > step 1 succeeds. app crashes. step 2 never runs. > data is now inconsistent permanently 16/ health check endpoint missing > load balancer sends traffic to crashed instances > users get 502s. you get support emails. 17/ memory leaks in long-running processes > memory grows slowly. server hits 100%. > everything grinds to a halt. restart. repeat. 18/ no graceful shutdown handling > deploy kills active requests > users mid-action get errors with no retry 19/ dependent on a third-party API with no fallback > that API goes down > your core feature goes down with it 20/ all logs written to local disk > logs rotate off > incident happens. no history to debug with. 21/ no circuit breaker on external calls > external service is slow > your thread pool fills waiting for it. everything queues. 22/ unparameterized search queries > search with any real data volume > 5-second response times at scale 23/ no connection timeout on outbound HTTP calls > external API hangs > your thread hangs with it. indefinitely. 24/ WebSockets not handled by a stateful service > horizontal scale breaks real-time features > every user gets disconnected 25/ no runbook for common incidents > something breaks at 2 AM > nobody knows what to do. everyone panics. you can ship fast and still build something that holds. bookmark this before you go live.

@halhjri14 بالتوفيق ان شاءالله. Love to help if possible

نفتخر في سنونو ان عندنا أبطال من العراق العظيم مهد الحضارات ، وأولهم الاخ مثنى الفارس 🔥🇮🇶

@halhjri14 @salesforce You’re spot on. A lot of platforms become bloated beyond what businesses here really need. There’s definitely a gap for something more practical and region-focused.

يستخدم العالم العربي أحد افشل التقنيات على مستوى العالم لادارة المبيعات ، الشركة الأمريكية @salesforce .. لماذا لايكون هناك شاب طموح يبني لنا شي بديل وكلنا راح ندعمه .. أنا أحد اول المستثمرين ..

That is what we need to do in Saudi Arabia and GCC countries

@alghali أخي ثامر من الضروري ان نطور مثل هذا التكنولوجيا في بلدننا (بلاد الخليج) . أنا متأكد من إمكانيه ذالك.

🚨🇫🇷 عاجل شركة Aerix Systems الفرنسية تكشف عن طائرة مسيّرة تحافظ على استقرارها حتى بعد تعطّل أحد المراوح في الهواء! 🛸 طائرة AXS-μ1 مصمّمة للاستمرار في الطيران في مواقف كانت تؤدي حتماً إلى تحطّم أي درون آخر. إذا نجحت هذه التقنية، ستُعيد رسم قواعد اللعبة في: 🔹 المراقبة 🔹 الدفاع 🔹 البيئات عالية الخطورة المرونة لم تعد ميزة إضافية، أصبحت ضرورة. ✍️ Defence_Index

@abidalista @POGYaz Thanks. Sincerely and from my heart I would love to help and contribute there 🙏🏻

@POGYaz عندك هنا اسماء شركات منقّاة تساهم في بناء المستقبل في السعودية. بالتوفيق buildsaudi.co

بحكم ان الفترة القادمة بتكون فترة Internship للأغلب.. اذكروا لنا تحت هذي التغريده "شركات\جهات" ودكم تتدربوا فيها 👀! وليش؟

@malmadhi2018 @ycombinator بالفعل، بنينا شيئًا مشابهًا قبل كم سنة. التحدي لم يكن في النموذج نفسه، لأن النتائج كانت جيدة جدًا، بل في الوصول إلى البيانات وإقناع الفرق القانونية باستخدام النظام. في المجال القانوني تحديدًا، الحساسية العالية تجاه الخصوصية والسرية تجعل تبني هذه الحلول أصعب من بناء التقنية نفسها.

شركة Legora قانونية تستخدم الذكاء الاصطناعي نموذج عملها يشبه شركة Harvey الشركة بدات في السويد عام 2023 وانتقلت الى أمريكا بعدها في خلال حوالي عامين وصل تقييمها الى نحو 5.5 مليار دولار دخلت مسرعة @ycombinator حققت حوالي مليون دولار كإيرادات في شهر ونصف فقط تقدم المنصة خدمات قانونية كاملة مثل: -المراجعة: تحليل آلاف الصفحات خلال دقائق - الصياغة: اقتراح عقود وبنود جاهزة داخل المنصة - البحث: ربط البيانات القانونية بالإضافة الى بيانات الشركة الفكرة هنا ليست في بناء chatbot قانوني، بل في إدخال الذكاء الاصطناعي مباشرة داخل سير العمل اليومي، بحيث يصبح جزءاً من المراجعة والصياغة والبحث والتعاون، لا طبقة منفصلة عنه. #الاسهم_الامريكية #استثمار youtu.be/I-BERqPprrc?si…