Shukhratbek.dev

Mentor sifatida yana faoliyatni qayta boshlayapman, hozircha kompaniyada Node.js bo'yicha, Global Mentoring Programga qo'shildim. Ko'p muddat buni pauzaga qo'ygan edim (2019 yildan boshlab). Keyinchalik Xudo hohlasa shu yerda ham boshlash niyat bor.

Hamkasblardan FrontEnd interview haqida talk bo'lar ekan FrontSpotda (Rus tilida). Online 13 Fevral soat 20:00 Toshkent vaqti bo'yicha. https://wearecommunity.io/events/frontspot-talks-front-end-intervyu-dostizhim-li-ideal

eCommercedagilarga zo'r taklifim bor, endi real professional photo sessiya keragi yo'q. O'zim zo'r qilib kiyimlarni AI modellarga kiydirib beraman 😀 tez, sifatli va qimmat 😎 AI shu soha ni ham egallamoqda ekan. Natijalar yomon emas.

Diplom project uchun mavzu tanladim: AI-Powered Health Dashboard: An Automated System for Medical Test Analysis and Personalized Recommendations This project introduces an AI-powered health dashboard designed to empower users with actionable insights derived from their medical test results. By leveraging cutting-edge technologies such as Optical Character Recognition (OCR), machine learning, and data visualization, the system parses medical reports (e.g., blood, urine, liver tests) to identify critical health markers, evaluate trends, and provide evidence-based recommendations. The platform further enhances user value by suggesting relevant vitamins, supplements, and lifestyle changes based on deficiencies or identified risks, with potential product links (e.g., iHerb) for added convenience. The project bridges the gap between raw medical data and actionable knowledge, helping users monitor their health, predict risks, and make informed decisions about their well-being, thus increasing early awareness and promoting healthier lifestyles. O'ylab-o'ylab shu mazvuni tanladim, life-science va AI(xar xil LLM bilan ishlash) juda qiziq, avvalo o'zim uchun bu narsa. yaxshi natija chiqsa startup qivoraman 😊

State management uchun Zustand ni sinab ko'rgandim bitta loyihada. Juda sodda yechim edi Redux o'rniga. Yaqinda yana bitta lib ni topib oldim. effector - https://effector.dev/ uni ham bir sinab ko'rish kerak.

AI ga qiziqqanlar uchun. Bepul NVIDIA Generative AI Course (pullik kursdan bittasini tanlash imkoniyati) Bu kurslar $90 gacha, ulardan birini bepul tanlab o'qish imkoniyati berilibdi. Kurslar self-paced va tugatilgandan so'ng certificate berishar ekan: - Exploring Adversarial Machine Learning - Fundamentals of Accelerated Computing with CUDA Python - Generative AI with Diffusion Models - Getting Started with Accelerated Computing in CUDA C/C++ - Getting Started with Deep Learning - Introduction to Deploying RAG Pipelines for Production at Scale - Introduction to Transformer-Based Natural Language Processing - Sizing LLM Inference Systems - Techniques for Improving the Effectiveness of RAG Systems https://sp-events.courses.nvidia.com/dli-india25?ncid=ref-inpa-419622 NVIDIA Deep Learning Institute Enrollment Benefits

Meetinglar(remote) uchun foydali sherik. Bizning kasbni eng asosiy muammosi bu ko'p o'tirish va vazn olish. Shuni bir sinab ko'rmoqchiman ham gaplashib, ham yurib. O'ylab qarasam o'rtacha har kuni kamida 2 soat meetingim bor ekan. Foyda berar 🙃

Yaqinda loyihada bir improvement qilgan edim. Ishimizni supportda osonlashtirgani. Ya'ni muammo - foydalanuvchida qanaqadur xato yuzaga kelsa uning xatosini millionta logni ichidan top o'sha foydalanuvchiga tegishligini, agar unda bir nechta xato bo'lsa qaysi haqida gap ketyapbi undan so'ra, muammoni aniqla va uni bartaraf qilish ko'p vaqt olayotgan edi. Oldin error toast to'liq errorni ko'rsatar edi, bu foydalanuvchiga umuman qiziq emas. va foydalanuvchu emailni orqali muammoni o'z tilida yozar edi yoki screenshot tashar edi. Bu unchalik informative emas va analiz qilish uchun ko'p vaqt talab qilar edi. Yechim oddiy bo'ldi. error toast messageni ixchamlashtirildi va unga correlationId(AWS request ID) qo'shildi va danger button qilindi. Natijada foydalanuvchi faqat qizil(danger button) tugamsini bosishi kifoya, unda analiz uchun kerakli ma'lumotlar qo'shilgan bo'ladi mailto orqali va u bizga shu emailni jo'natadi. Va o'sha correlationID orqali microserviceda muammoni oson topiladi. Ha bu masalani yechish uchun yo'llar ko'p, lekin bu bizning processlar uchun eng oson va unumdori deb topildi 😊

Loyihani rivolanish davrida Documentationni up-to-date saqlash eng katta muammolardan biri bo'lsa kerak, ayniqsa qaysidur bir module ning logic qismi. Misol uchun biz Documentation uchun source-of-truth sifatida KB(Confluence) ni ishlatamiz. Lekin function/component/module logikasi o'zgargandan so'ng documentation ni yangilash esdan chiqishi mumkin developer tomonidan. Va bizda keyin KBda document out-dated ga aylanadi. Buni up-to-date saqlash ham bir discipline. Bu masalani yechish uchun - code ni o'ziga tsdoc va typedoc yordamida mermaid DaC(Diagram as Code) orqali diagram larni qo'shishni experiment qilyapman. Keyingi step CI qismida o'zgaragan diagramni KBga publish qilish bo'ladi 😊 Developer changes logic -> updates comment doc/diagram by herself/himself or uses AI -> pass Code Review -> deliver updated document to target. Maqsad iloji boricha ishni automate qilish va developer ishini osonlashtirib, human-factor riskni oldini olish.

AI Engineer vacancy lar ham chiqibti. Men ishlaydigan kompaniya hozirda Senior/Lead AI Engineer qidiryapti O'zbekistonda, agarda bilimingiz mos kelsa referral qilishga yordam berishim mumkin. Responsibilities Design and implement advanced solutions for complex challenges in GenAI Develop, refine and optimize AI systems within our in-house products, enhancing better SDLC Engage in client-specific projects to integrate AI solutions and address specific business challenges Resolve unpredictable challenges across technical and infrastructural domains Adapt new technologies and methodologies to maintain cutting-edge development practices Requirements Proven expertise in software development at a Senior level up to an Architect role Proficiency in Python with a focus on Python frameworks in GenAI Skills in other programming languages such as Java, .NET, C/C++ or JavaScript Keen interest in GenAI for continuous professional growth within this domain Motivation to be hands-on in both coding and implementation, beyond just architectural design Enthusiasm for both in-house product development and external projects with AI integration Nice to have Experience in production environments within GenAI

Diploma project boshlashni vaqti ham kelibdi. Nima qilsam ekan? qaysi topic ni tanlasam ekan? degan savollar qiynayapti endi. Univer tomonidan tayyor Research Advisor lar beiladi, research muammo bo'lmaydi deb umid qilaman. Oddiy CRUD servis qilsam osonuyat bo'lar 😅 Hozir AI(SI) trendda, undan 10500% foydalanish kerak. Balkim Maktab AI qilsammikan 😀 - mendaqa ota-onalarga bolani o'quv processini track qilgani va yo'naltirgani.

Siz o'zi nima qilasiz ishda deb so'rab qolishadi ba'zida. Code yozasizmi xolos deb so'rashadi. 2024 yilga shunaqa natija ekan code yozish bo'yicha. Code yozish unchalik ko'p vaqt olmaydi meni ishimda, asosiysi bu Spike/Research, RFC, ADR, System Design/Sketch Diagrams va Architecture Documentionlar yaratish va ularni jamoa bn tasdiqlash, undan keyingina code yozishga kirishiladi. Yana eng asosiy qismi bu meetinglar - haftasiga 10-15 soat meetinglarga sarflar ekanman: ularni ichida daily standup, refinement, planning, syncup va h.z.

Xudo hohlasa Contentstackdan Certificate olishga harakatni boshladim. Uni 2-3 yildan beri ishlataman proyektlarda, shuning uchun ozgina bo'lsa ham malaka bor endi. Lekin Headless va Composable Architecture juda qiziq va boshaqcha fikrlashga undaydi.

Hozircha yaxshi TTS (Text-to-speech) AI topolmadim o'zbekchada o'zimni ovozim orqali generatsiya qila oladigan servis yoki AI ni. O'zimiznikilar yaratganlar esa - huddi oldingi Googlenikiga o'xshaydi talaffuzi, xuddi robot gapirganday. ElevenLabs zo'r lekin unda O'zbekcha yo'q. Turkcha qilib yuboryapti 😀 Agar kimdir bilsa, share qivorilar iltimos.

AI Automation da ilk qadam. n8n yordamida faceless Youtube shorts yaratildi. Ovoz ham o'zimnikidan clone qilingan. Endi Promptlarni optimizatsiya qilish kerak.

PayPal haliyam O'zbekiston borligini bilmas ekan 🤬 Deepseek ga pul to'lash imkoniyati ham qolmabdi

Hash-pash deguncha 9 yil ham bo'libdi EPAMda ishlayotganimga 🎉🙃

qiziq mock-interview ekan DSA - Data Structures & Algorithms Interview https://youtu.be/6ncshuA-jqo?si=eBLyylIsI9G2bCVz

Bugun Happy Programmer's Day Bayram muborak bo'lsin 🎉 Bugun ortiqcha kofe bn nishonlab ☕️ qo'yamiz https://www.timeanddate.com/holidays/world/international-programmers-day