{"id":36571,"date":"2025-12-20T02:00:36","date_gmt":"2025-12-19T23:00:36","guid":{"rendered":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/"},"modified":"2025-12-20T02:00:36","modified_gmt":"2025-12-19T23:00:36","slug":"bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik","status":"publish","type":"post","link":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/","title":{"rendered":"Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k"},"content":{"rendered":"<h2>Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k<\/h2>\n<p>Merhaba de\u011ferli builder&#8217;lar! Makine \u00f6\u011frenimi (ML) d\u00fcnyas\u0131na ad\u0131m atmak, \u00f6zellikle AWS gibi kapsaml\u0131 bir bulut platformunda, ilk ba\u015fta g\u00f6z korkutucu g\u00f6r\u00fcnebilir. Ancak endi\u015felenmeyin, bu yolculukta yaln\u0131z de\u011filsiniz. Bu makale, s\u0131f\u0131rdan ba\u015flayarak AWS Makine \u00d6\u011frenimi ekosistemine sa\u011flam bir giri\u015f yapman\u0131z i\u00e7in bir rehber niteli\u011findedir. Temel kavramlardan pratik uygulamalara kadar ad\u0131m ad\u0131m ilerleyece\u011fiz, b\u00f6ylece kendi ML projelerinizi hayata ge\u00e7irecek donan\u0131ma sahip olacaks\u0131n\u0131z.<\/p>\n<h3>Yolculu\u011fa Ba\u015flarken: Temel Kavramlar\u0131 Anlamak<\/h3>\n<p>Makine \u00f6\u011frenimi ser\u00fcvenine \u00e7\u0131kmadan \u00f6nce, temel ta\u015flar\u0131 do\u011fru bir \u015fekilde yerle\u015ftirmek b\u00fcy\u00fck \u00f6nem ta\u015f\u0131r. Bu b\u00f6l\u00fcmde, makine \u00f6\u011freniminin ne oldu\u011funu, neden AWS&#8217;i tercih etmemiz gerekti\u011fini ve bu alandaki temel terimleri ele alaca\u011f\u0131z.<\/p>\n<h4>Makine \u00d6\u011frenimi Nedir? Neden AWS?<\/h4>\n<p>Makine \u00f6\u011frenimi, bilgisayar sistemlerinin a\u00e7\u0131k\u00e7a programlanmadan verilerden \u00f6\u011frenmesini sa\u011flayan yapay zeka (YZ) alt dal\u0131d\u0131r. K\u0131sacas\u0131, makinelerin deneyimlerden ders \u00e7\u0131kararak gelecekteki kararlar\u0131 daha iyi almas\u0131n\u0131 sa\u011flar. Peki, neden AWS?<\/p>\n<ul>\n<li><strong>\u00d6l\u00e7eklenebilirlik:<\/strong> AWS, k\u00fc\u00e7\u00fck projelerden b\u00fcy\u00fck \u00f6l\u00e7ekli kurumsal \u00e7\u00f6z\u00fcmlere kadar her t\u00fcrl\u00fc ML i\u015f y\u00fck\u00fcn\u00fc kolayca y\u00f6netmenizi sa\u011flar.<\/li>\n<li><strong>Kapsaml\u0131 Hizmetler:<\/strong> SageMaker gibi \u00f6zel ML hizmetlerinden S3 gibi depolama \u00e7\u00f6z\u00fcmlerine, EC2 gibi i\u015flem g\u00fcc\u00fcne kadar geni\u015f bir hizmet yelpazesi sunar.<\/li>\n<li><strong>Maliyet Etkinli\u011fi:<\/strong> Kulland\u0131k\u00e7a \u00f6de modeli sayesinde sadece ihtiya\u00e7 duydu\u011funuz kaynaklar i\u00e7in \u00f6deme yapars\u0131n\u0131z.<\/li>\n<li><strong>Yenilik:<\/strong> AWS, s\u00fcrekli olarak yeni ML hizmetleri ve \u00f6zellikleri sunarak sekt\u00f6rdeki lider konumunu korur.<\/li>\n<\/ul>\n<h4>Temel ML Terimleri ve Algoritmalara Genel Bak\u0131\u015f<\/h4>\n<p>ML d\u00fcnyas\u0131nda s\u0131k\u00e7a kar\u015f\u0131la\u015faca\u011f\u0131n\u0131z baz\u0131 temel terimler \u015funlard\u0131r:<\/p>\n<ul>\n<li><strong>Model:<\/strong> Verilerden \u00f6\u011frenen ve tahminler yapan matematiksel yap\u0131.<\/li>\n<li><strong>E\u011fitim (Training):<\/strong> Modelin veri k\u00fcmesi \u00fczerinde \u00f6\u011frenme s\u00fcreci.<\/li>\n<li><strong>\u00c7\u0131kar\u0131m (Inference):<\/strong> E\u011fitilmi\u015f modelin yeni veriler \u00fczerinde tahmin yapmas\u0131.<\/li>\n<li><strong>Denetimli \u00d6\u011frenme (Supervised Learning):<\/strong> Etiketli verilerle (girdi-\u00e7\u0131kt\u0131 \u00e7iftleri) modelin e\u011fitilmesi. Regresyon ve s\u0131n\u0131fland\u0131rma en yayg\u0131n t\u00fcrleridir.<\/li>\n<li><strong>Denetimsiz \u00d6\u011frenme (Unsupervised Learning):<\/strong> Etiketsiz verilerle desenleri ve yap\u0131lar\u0131 ke\u015ffetme. K\u00fcmeleme ve boyut indirgeme \u00f6rnekleridir.<\/li>\n<li><strong>Takviyeli \u00d6\u011frenme (Reinforcement Learning):<\/strong> Bir ajan\u0131n bir ortamda etkile\u015fim kurarak \u00f6d\u00fcl ve ceza mekanizmas\u0131yla \u00f6\u011frenmesi.<\/li>\n<\/ul>\n<p>Pop\u00fcler algoritmalardan baz\u0131lar\u0131:<\/p>\n<ul>\n<li><strong>Lineer Regresyon:<\/strong> S\u00fcrekli de\u011ferleri tahmin etmek i\u00e7in.<\/li>\n<li><strong>Lojistik Regresyon:<\/strong> \u0130kili s\u0131n\u0131fland\u0131rma i\u00e7in.<\/li>\n<li><strong>Karar A\u011fa\u00e7lar\u0131\/Rastgele Ormanlar:<\/strong> Hem regresyon hem de s\u0131n\u0131fland\u0131rma i\u00e7in g\u00fc\u00e7l\u00fc ve yorumlanabilir modeller.<\/li>\n<li><strong>Destek Vekt\u00f6r Makineleri (SVM):<\/strong> S\u0131n\u0131fland\u0131rma ve regresyon i\u00e7in.<\/li>\n<li><strong>K-Means:<\/strong> K\u00fcmeleme i\u00e7in.<\/li>\n<\/ul>\n<h4>Veri Bilimi ve M\u00fchendisli\u011fi Rolleri<\/h4>\n<p>ML projelerinde genellikle iki ana rol \u00f6ne \u00e7\u0131kar:<\/p>\n<ul>\n<li><strong>Veri Bilimcisi:<\/strong> Verileri analiz eder, modeller geli\u015ftirir ve i\u015f problemlerine \u00e7\u00f6z\u00fcmler \u00fcretir. \u0130statistik, matematik ve programlama bilgisi g\u00fc\u00e7l\u00fcd\u00fcr.<\/li>\n<li><strong>Makine \u00d6\u011frenimi M\u00fchendisi:<\/strong> Geli\u015ftirilen modelleri \u00fcretim ortam\u0131na ta\u015f\u0131r, \u00f6l\u00e7eklenebilir altyap\u0131lar kurar ve model performans\u0131n\u0131 izler. Yaz\u0131l\u0131m m\u00fchendisli\u011fi ve da\u011f\u0131t\u0131k sistemler bilgisi \u00f6n plandad\u0131r.<\/li>\n<\/ul>\n<p>Bu yolculukta, her iki rol\u00fcn de temel yetkinliklerine dokunaca\u011f\u0131z.<\/p>\n<h3>AWS Ortam\u0131n\u0131 Tan\u0131ma: Hesap Kurulumu ve G\u00fcvenlik<\/h3>\n<p>AWS&#8217;te ML projelerine ba\u015flamadan \u00f6nce, g\u00fcvenli ve d\u00fczenli bir \u00e7al\u0131\u015fma ortam\u0131 olu\u015fturmak kritik \u00f6neme sahiptir. Bu b\u00f6l\u00fcmde, AWS hesab\u0131 olu\u015fturmaktan maliyet y\u00f6netimine kadar temel ad\u0131mlar\u0131 ele alaca\u011f\u0131z.<\/p>\n<h4>AWS Hesab\u0131 Olu\u015fturma ve Temel Ayarlar<\/h4>\n<p>AWS&#8217;e ba\u015flamak i\u00e7in \u00f6ncelikle bir hesap olu\u015fturman\u0131z gerekir. Bu s\u00fcre\u00e7 olduk\u00e7a basittir ve kredi kart\u0131 bilgilerinizi gerektirir (\u00fccretsiz katman avantajlar\u0131ndan yararlanmak i\u00e7in bile). Hesab\u0131n\u0131z\u0131 olu\u015fturduktan sonra, ilk yapman\u0131z gerekenler:<\/p>\n<ul>\n<li><strong>Root Kullan\u0131c\u0131 Parolas\u0131n\u0131 G\u00fc\u00e7lendirme:<\/strong> Karma\u015f\u0131k bir parola belirleyin ve MFA (Multi-Factor Authentication) etkinle\u015ftirin.<\/li>\n<li><strong>B\u00f6lge Se\u00e7imi:<\/strong> Projenizin co\u011frafi konumuna en yak\u0131n AWS b\u00f6lgesini se\u00e7mek, gecikmeyi azalt\u0131r ve bazen maliyetleri d\u00fc\u015f\u00fcr\u00fcr.<\/li>\n<\/ul>\n<h4>IAM ile G\u00fcvenli Eri\u015fim Y\u00f6netimi<\/h4>\n<p>Root kullan\u0131c\u0131n\u0131zla g\u00fcnl\u00fck i\u015flerinizi yapmaktan ka\u00e7\u0131nmal\u0131s\u0131n\u0131z. Bunun yerine, AWS Identity and Access Management (IAM) kullanarak farkl\u0131 kullan\u0131c\u0131lar, gruplar ve roller olu\u015fturmal\u0131s\u0131n\u0131z. Bu, yetkilendirmeyi gran\u00fcler bir \u015fekilde y\u00f6netmenizi sa\u011flar.<\/p>\n<pre><code class=\"language-json\">{\n    \"Version\": \"2012-10-17\",\n    \"Statement\": [\n        {\n            \"Effect\": \"Allow\",\n            \"Action\": [\n                \"s3:GetObject\",\n                \"s3:PutObject\"\n            ],\n            \"Resource\": \"arn:aws:s3:::your-ml-bucket\/*\"\n        },\n        {\n            \"Effect\": \"Allow\",\n            \"Action\": \"sagemaker:*\",\n            \"Resource\": \"*\"\n        }\n    ]\n}<\/pre>\n<p><\/code><\/p>\n<p>Yukar\u0131daki \u00f6rnek bir IAM politikas\u0131, belirli bir S3 kovas\u0131na eri\u015fim ve t\u00fcm SageMaker i\u015flemlerini yapma yetkisi verir. Her kullan\u0131c\u0131ya veya role sadece ihtiya\u00e7 duydu\u011fu izinleri atamak, \"en az ayr\u0131cal\u0131k\" prensibinin temelidir.<\/p>\n<h4>Maliyet Y\u00f6netimi ve B\u00fct\u00e7e Kontrol\u00fc<\/h4>\n<p>AWS'te maliyetler h\u0131zla artabilir. Bunu \u00f6nlemek i\u00e7in:<\/p>\n<ul>\n<li><strong>B\u00fct\u00e7eler Olu\u015fturma:<\/strong> AWS B\u00fct\u00e7eler servisini kullanarak belirli bir e\u015fi\u011fe ula\u015f\u0131ld\u0131\u011f\u0131nda uyar\u0131lar alabilirsiniz.<\/li>\n<li><strong>Tagging (Etiketleme):<\/strong> Kaynaklar\u0131n\u0131za (EC2, S3, SageMaker) proje, departman gibi etiketler atayarak maliyetleri daha iyi takip edebilirsiniz.<\/li>\n<li><strong>Kullan\u0131lmayan Kaynaklar\u0131 Kapatma:<\/strong> \u00d6zellikle geli\u015ftirme a\u015famas\u0131nda, kullanmad\u0131\u011f\u0131n\u0131z SageMaker notebook'lar\u0131n\u0131 veya EC2 instance'lar\u0131n\u0131 kapatmay\u0131 unutmay\u0131n.<\/li>\n<\/ul>\n<h3>Veri Haz\u0131rl\u0131\u011f\u0131: ML'in Yak\u0131t\u0131<\/h3>\n<p>Makine \u00f6\u011freniminde \"\u00e7\u00f6p girdi, \u00e7\u00f6p \u00e7\u0131kt\u0131\" (garbage in, garbage out) diye bir s\u00f6z vard\u0131r. Modelinizin ba\u015far\u0131s\u0131 b\u00fcy\u00fck \u00f6l\u00e7\u00fcde verilerinizin kalitesine ba\u011fl\u0131d\u0131r. Bu b\u00f6l\u00fcm, veri toplama, depolama ve \u00f6n i\u015fleme ad\u0131mlar\u0131n\u0131 kapsar.<\/p>\n<h4>Veri Toplama ve Depolama (S3, RDS)<\/h4>\n<p>Verileriniz farkl\u0131 kaynaklardan gelebilir:<\/p>\n<ul>\n<li><strong>Amazon S3 (Simple Storage Service):<\/strong> B\u00fcy\u00fck miktarda yap\u0131land\u0131r\u0131lmam\u0131\u015f veri (resimler, videolar, metin dosyalar\u0131, CSV'ler) depolamak i\u00e7in idealdir. Dayan\u0131kl\u0131, \u00f6l\u00e7eklenebilir ve uygun maliyetlidir.<\/li>\n<li><strong>Amazon RDS (Relational Database Service):<\/strong> Yap\u0131land\u0131r\u0131lm\u0131\u015f veriler i\u00e7in (SQL veritabanlar\u0131). PostgreSQL, MySQL, SQL Server gibi se\u00e7enekler sunar.<\/li>\n<li><strong>Amazon DynamoDB:<\/strong> NoSQL veritaban\u0131, y\u00fcksek performansl\u0131 ve \u00f6l\u00e7eklenebilir anahtar-de\u011fer veya belge tabanl\u0131 veriler i\u00e7in.<\/li>\n<\/ul>\n<p>Genellikle ML projelerinde S3, veri g\u00f6l\u00fc (data lake) olarak kullan\u0131l\u0131r.<\/p>\n<h4>Veri Ke\u015ffi ve \u00d6n \u0130\u015fleme (Pandas, Glue)<\/h4>\n<p>Verileri toplad\u0131ktan sonra, onlar\u0131 anlamak ve ML modeline uygun hale getirmek gerekir:<\/p>\n<ul>\n<li><strong>Veri Ke\u015ffi (Exploratory Data Analysis - EDA):<\/strong> Veri setinizin yap\u0131s\u0131n\u0131, da\u011f\u0131l\u0131m\u0131n\u0131, eksik de\u011ferlerini ve ayk\u0131r\u0131 de\u011ferlerini anlamak i\u00e7in g\u00f6rselle\u015ftirme ve istatistiksel analizler yap\u0131l\u0131r (Python'da Pandas, Matplotlib, Seaborn k\u00fct\u00fcphaneleriyle).<\/li>\n<li><strong>Eksik De\u011ferleri Y\u00f6netme:<\/strong> Eksik verileri doldurma (ortalama, medyan, mod ile) veya ilgili sat\u0131rlar\u0131\/s\u00fctunlar\u0131 silme.<\/li>\n<li><strong>Ayk\u0131r\u0131 De\u011ferleri \u0130\u015fleme:<\/strong> Ayk\u0131r\u0131 de\u011ferleri tespit etme ve d\u00fczeltme veya kald\u0131rma.<\/li>\n<li><strong>Veri D\u00f6n\u00fc\u015ft\u00fcrme:<\/strong> Kategorik verileri say\u0131sal formata d\u00f6n\u00fc\u015ft\u00fcrme (One-Hot Encoding, Label Encoding). Say\u0131sal verileri normalle\u015ftirme veya standartla\u015ft\u0131rma.<\/li>\n<\/ul>\n<p>AWS Glue, sunucusuz bir veri entegrasyon hizmetidir. ETL (Extract, Transform, Load) i\u015f y\u00fckleri i\u00e7in kullan\u0131labilir, b\u00fcy\u00fck veri k\u00fcmelerini \u00f6n i\u015flemek i\u00e7in idealdir.<\/p>\n<h4>\u00d6zellik M\u00fchendisli\u011fi ve Veri B\u00f6l\u00fcmleme<\/h4>\n<ul>\n<li><strong>\u00d6zellik M\u00fchendisli\u011fi (Feature Engineering):<\/strong> Mevcut verilerden yeni, daha anlaml\u0131 \u00f6zellikler t\u00fcretme sanat\u0131d\u0131r. \u00d6rne\u011fin, tarih bilgisinden y\u0131l, ay, g\u00fcn veya haftan\u0131n g\u00fcn\u00fc gibi yeni \u00f6zellikler \u00e7\u0131kar\u0131labilir. Bu, modelin performans\u0131n\u0131 \u00f6nemli \u00f6l\u00e7\u00fcde art\u0131rabilir.<\/li>\n<li><strong>Veri B\u00f6l\u00fcmleme:<\/strong> Modelin genelleme yetene\u011fini de\u011ferlendirmek i\u00e7in veri setini genellikle \u00fc\u00e7 par\u00e7aya ay\u0131r\u0131r\u0131z:\n<ul>\n<li><strong>E\u011fitim Seti (Training Set):<\/strong> Modelin \u00f6\u011frenmesi i\u00e7in kullan\u0131l\u0131r.<\/li>\n<li><strong>Do\u011frulama Seti (Validation Set):<\/strong> Modelin hiperparametrelerini ayarlamak ve a\u015f\u0131r\u0131 \u00f6\u011frenmeyi (overfitting) \u00f6nlemek i\u00e7in kullan\u0131l\u0131r.<\/li>\n<li><strong>Test Seti (Test Set):<\/strong> Modelin son performans\u0131n\u0131, daha \u00f6nce g\u00f6rmedi\u011fi veriler \u00fczerinde de\u011ferlendirmek i\u00e7in kullan\u0131l\u0131r.<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>AWS SageMaker ile Tan\u0131\u015fma: ML Geli\u015ftirme Ortam\u0131<\/h3>\n<p>AWS SageMaker, makine \u00f6\u011frenimi modellerini olu\u015fturma, e\u011fitme ve da\u011f\u0131tma s\u00fcrecini basitle\u015ftiren tam y\u00f6netilen bir hizmettir. ML ya\u015fam d\u00f6ng\u00fcs\u00fcn\u00fcn her a\u015famas\u0131n\u0131 kapsar.<\/p>\n<h4>SageMaker Nedir? Bile\u015fenleri Nelerdir?<\/h4>\n<p>SageMaker, ML uzmanlar\u0131n\u0131n ve geli\u015ftiricilerin modelleri daha h\u0131zl\u0131 bir \u015fekilde \u00fcretime ge\u00e7irmelerine yard\u0131mc\u0131 olmak i\u00e7in tasarlanm\u0131\u015ft\u0131r. Temel bile\u015fenleri:<\/p>\n<ul>\n<li><strong>SageMaker Notebook Instance'lar:<\/strong> Etkile\u015fimli geli\u015ftirme ortamlar\u0131 (Jupyter Notebook\/Lab).<\/li>\n<li><strong>SageMaker Training:<\/strong> Model e\u011fitim i\u015flerini y\u00f6netmek i\u00e7in \u00f6l\u00e7eklenebilir altyap\u0131.<\/li>\n<li><strong>SageMaker Hosting (Endpoints):<\/strong> E\u011fitilmi\u015f modelleri ger\u00e7ek zamanl\u0131 \u00e7\u0131kar\u0131m i\u00e7in da\u011f\u0131tma.<\/li>\n<li><strong>SageMaker Processing:<\/strong> B\u00fcy\u00fck \u00f6l\u00e7ekli veri \u00f6n i\u015fleme, \u00f6zellik m\u00fchendisli\u011fi ve model de\u011ferlendirme i\u015fleri.<\/li>\n<li><strong>SageMaker Pipelines:<\/strong> ML i\u015f ak\u0131\u015flar\u0131n\u0131 otomatikle\u015ftirmek i\u00e7in CI\/CD.<\/li>\n<\/ul>\n<h4>Notebook Instance'lar ve Geli\u015ftirme Ortam\u0131<\/h4>\n<p>SageMaker Notebook Instance'lar\u0131, ML geli\u015ftirme s\u00fcrecinin kalbidir. Python, R, Julia gibi diller ve pop\u00fcler ML k\u00fct\u00fcphaneleri (TensorFlow, PyTorch, Scikit-learn) \u00f6nceden y\u00fcklenmi\u015f olarak gelir. Bir notebook instance ba\u015flatmak olduk\u00e7a kolayd\u0131r:<\/p>\n<ol>\n<li>AWS Y\u00f6netim Konsolu'nda SageMaker hizmetine gidin.<\/li>\n<li>\"Notebook instances\" alt\u0131nda \"Create notebook instance\" se\u00e7ene\u011fini t\u0131klay\u0131n.<\/li>\n<li>Bir isim verin, instance tipini se\u00e7in (\u00f6rne\u011fin, ml.t2.medium) ve bir IAM rol\u00fc atay\u0131n.<\/li>\n<li>Olu\u015fturduktan sonra \"Open Jupyter\" veya \"Open JupyterLab\" ile ortam\u0131n\u0131za eri\u015febilirsiniz.<\/li>\n<\/ol>\n<pre><code class=\"language-python\">import sagemaker\nfrom sagemaker.pytorch import PyTorch\n\n# SageMaker oturumu ba\u015flat\nsagemaker_session = sagemaker.Session()\n\n# PyTorch estimator tan\u0131mlama\nestimator = PyTorch(\n    entry_point='train.py',\n    role=sagemaker.get_execution_role(),\n    framework_version='1.10.0',\n    py_version='py38',\n    instance_count=1,\n    instance_type='ml.m5.xlarge',\n    hyperparameters={\n        'epochs': 10,\n        'batch-size': 64\n    }\n)\n\n# Modeli e\u011fitme\nestimator.fit({'training': 's3:\/\/your-s3-bucket\/data'})<\/pre>\n<p><\/code><\/p>\n<p>Bu Python kodu, SageMaker SDK kullanarak bir PyTorch modelini e\u011fitmek i\u00e7in bir estimator tan\u0131mlar ve e\u011fitim i\u015fini ba\u015flat\u0131r.<\/p>\n<h4>Modelleri E\u011fitme ve Ayarlama (Training Jobs, Hyperparameter Tuning)<\/h4>\n<p>SageMaker, model e\u011fitimini \u00f6l\u00e7eklenebilir ve y\u00f6netilebilir hale getirir. E\u011fitim i\u015flerinizi (Training Jobs) farkl\u0131 instance tipleri (CPU veya GPU) \u00fczerinde \u00e7al\u0131\u015ft\u0131rabilirsiniz. B\u00fcy\u00fck veri k\u00fcmeleri veya karma\u015f\u0131k modeller i\u00e7in da\u011f\u0131t\u0131k e\u011fitimi destekler.<\/p>\n<p>Model performans\u0131n\u0131 optimize etmek i\u00e7in hiperparametre ayarlamas\u0131 (Hyperparameter Tuning) kritiktir. SageMaker Hyperparameter Tuning, belirli bir aral\u0131ktaki hiperparametre kombinasyonlar\u0131n\u0131 otomatik olarak test ederek en iyi performans\u0131 veren kombinasyonu bulman\u0131za yard\u0131mc\u0131 olur. Bu, manuel deneme yan\u0131lma s\u00fcrecini ortadan kald\u0131r\u0131r.<\/p>\n<h3>Modelleri Da\u011f\u0131tma ve \u0130zleme<\/h3>\n<p>Bir ML modelini e\u011fitmek sadece ba\u015flang\u0131\u00e7t\u0131r. Ger\u00e7ek de\u011fer, modelin \u00fcretim ortam\u0131nda kullan\u0131ma sunulmas\u0131 ve performans\u0131n\u0131n s\u00fcrekli izlenmesidir.<\/p>\n<h4>SageMaker Endpoint'ler ile Model Da\u011f\u0131t\u0131m\u0131<\/h4>\n<p>E\u011fitilmi\u015f bir modeli ger\u00e7ek zamanl\u0131 tahminler yapmak \u00fczere kullan\u0131ma sunmak i\u00e7in SageMaker Endpoint'leri kullan\u0131l\u0131r. Endpoint'ler, modelinizi bar\u0131nd\u0131ran ve gelen \u00e7\u0131kar\u0131m isteklerini i\u015fleyen tam y\u00f6netilen bir API sa\u011flar. Modelinizi da\u011f\u0131tmak genellikle tek bir kod sat\u0131r\u0131 ile yap\u0131labilir:<\/p>\n<pre><code class=\"language-python\">predictor = estimator.deploy(\n    initial_instance_count=1,\n    instance_type='ml.m5.xlarge'\n)\n\n# Tahmin yapma\nresponse = predictor.predict(test_data)\nprint(response)\n\n# Endpoint'i silme (maliyet kontrol\u00fc i\u00e7in \u00f6nemli!)\npredictor.delete_endpoint()<\/pre>\n<p><\/code><\/p>\n<p>Bu kod, e\u011fitilmi\u015f modeli bir SageMaker Endpoint olarak da\u011f\u0131t\u0131r ve ard\u0131ndan test verileri \u00fczerinde tahmin yapar. \u0130\u015finiz bitti\u011finde endpoint'i silmeyi unutmay\u0131n!<\/p>\n<h4>Modelleri \u0130zleme ve G\u00fcncelleme (Model Monitor)<\/h4>\n<p>\u00dcretimdeki modellerin zamanla performans\u0131 d\u00fc\u015febilir (model drift). SageMaker Model Monitor, da\u011f\u0131t\u0131lm\u0131\u015f modellerin performans\u0131n\u0131 ve veri kalitesini s\u00fcrekli olarak izlemenizi sa\u011flar. Veri kaymas\u0131 (data drift) veya kavram kaymas\u0131 (concept drift) tespit edildi\u011finde uyar\u0131lar g\u00f6nderir. Bu sayede modelinizi yeniden e\u011fitebilir veya g\u00fcncelleyebilirsiniz.<\/p>\n<p>Modelinizi g\u00fcncellemek i\u00e7in, yeni bir model s\u00fcr\u00fcm\u00fcn\u00fc e\u011fitip mevcut endpoint'i yeni modelle de\u011fi\u015ftirebilirsiniz (A\/B testleri veya mavi\/ye\u015fil da\u011f\u0131t\u0131m stratejileriyle).<\/p>\n<h4>Otomatik ML (AutoML) ile Tan\u0131\u015fma: SageMaker Autopilot<\/h4>\n<p>Makine \u00f6\u011frenimi uzmanl\u0131\u011f\u0131na sahip olmayan veya h\u0131zl\u0131 prototipleme yapmak isteyenler i\u00e7in SageMaker Autopilot harika bir \u00e7\u00f6z\u00fcmd\u00fcr. Autopilot, veri setinize g\u00f6re en iyi ML modelini, algoritmalar\u0131 ve hiperparametreleri otomatik olarak se\u00e7er ve e\u011fitir. Veri \u00f6n i\u015fleme, \u00f6zellik m\u00fchendisli\u011fi, model se\u00e7imi ve ayar\u0131n\u0131 sizin yerinize yapar.<\/p>\n<pre><code class=\"language-python\">from sagemaker.autopilot import AutopilotJob\n\nautopilot_job = AutopilotJob(\n    base_job_name='my-autopilot-job',\n    sagemaker_session=sagemaker_session,\n    target_attribute_name='target_column' # Tahmin edilecek s\u00fctun ad\u0131\n)\n\nautopilot_job.fit(\n    inputs='s3:\/\/your-s3-bucket\/data\/train.csv',\n    problem_type='BinaryClassification', # Regresyon, \u00c7oklu S\u0131n\u0131fland\u0131rma da olabilir\n    job_objective='F1' # Hedef metrik\n)<\/pre>\n<p><\/code><\/p>\n<p>Bu kod, Autopilot'u kullanarak otomatik bir ML i\u015fi ba\u015flat\u0131r. Bu, \u00f6zellikle h\u0131zl\u0131 ba\u015flang\u0131\u00e7lar i\u00e7in veya karma\u015f\u0131k model se\u00e7im s\u00fcre\u00e7lerinden ka\u00e7\u0131nmak istedi\u011finizde \u00e7ok faydal\u0131d\u0131r.<\/p>\n<h3>\u0130leri Konular ve Sonraki Ad\u0131mlar<\/h3>\n<p>Temelleri att\u0131ktan sonra, ML yolculu\u011funuzda ke\u015ffedebilece\u011finiz daha bir\u00e7ok ileri konu ve hizmet bulunmaktad\u0131r.<\/p>\n<h4>Derin \u00d6\u011frenme ve \u00d6zel Donan\u0131mlar (GPU'lar)<\/h4>\n<p>G\u00f6r\u00fcnt\u00fc i\u015fleme, do\u011fal dil i\u015fleme gibi alanlarda derin \u00f6\u011frenme (Deep Learning) modelleri (yapay sinir a\u011flar\u0131) \u00fcst\u00fcn performans g\u00f6sterir. Bu modeller genellikle \u00e7ok b\u00fcy\u00fck veri setleri \u00fczerinde e\u011fitilir ve yo\u011fun hesaplama g\u00fcc\u00fc gerektirir. AWS, \u00f6zellikle SageMaker ve EC2 hizmetleri arac\u0131l\u0131\u011f\u0131yla GPU destekli instance'lar sunarak bu t\u00fcr i\u015f y\u00fcklerini kolayca y\u00f6netmenizi sa\u011flar.<\/p>\n<h4>MLOps Kavram\u0131 ve S\u00fcre\u00e7leri<\/h4>\n<p>MLOps (Machine Learning Operations), yaz\u0131l\u0131m geli\u015ftirmedeki DevOps prensiplerini makine \u00f6\u011frenimi ya\u015fam d\u00f6ng\u00fcs\u00fcne uygulayan bir yakla\u015f\u0131md\u0131r. Ama\u00e7, ML modellerinin \u00fcretim ortam\u0131nda g\u00fcvenilir, verimli ve otomatik bir \u015fekilde da\u011f\u0131t\u0131lmas\u0131n\u0131, izlenmesini ve g\u00fcncellenmesini sa\u011flamakt\u0131r. SageMaker Pipelines, Model Registry ve Model Monitor gibi hizmetler, MLOps s\u00fcre\u00e7lerini uygulaman\u0131za yard\u0131mc\u0131 olur.<\/p>\n<h4>AWS ML Hizmet Ekosistemi (Rekognition, Comprehend, Textract vb.)<\/h4>\n<p>AWS, \u00f6zel ML uzmanl\u0131\u011f\u0131 gerektirmeyen, \u00f6nceden e\u011fitilmi\u015f bir\u00e7ok yapay zeka hizmeti sunar:<\/p>\n<ul>\n<li><strong>Amazon Rekognition:<\/strong> G\u00f6r\u00fcnt\u00fc ve video analizi (nesne tan\u0131ma, y\u00fcz analizi).<\/li>\n<li><strong>Amazon Comprehend:<\/strong> Do\u011fal dil i\u015fleme (duygu analizi, anahtar kelime \u00e7\u0131karma).<\/li>\n<li><strong>Amazon Textract:<\/strong> Belgelerden metin ve verileri otomatik olarak \u00e7\u0131karma.<\/li>\n<li><strong>Amazon Forecast:<\/strong> Zaman serisi tahminleri.<\/li>\n<li><strong>Amazon Personalize:<\/strong> Ki\u015fiselle\u015ftirilmi\u015f \u00f6neriler.<\/li>\n<\/ul>\n<p>Bu hizmetler, karma\u015f\u0131k ML modelleri geli\u015ftirmeden belirli yapay zeka yeteneklerini uygulamalar\u0131n\u0131za entegre etmenizi sa\u011flar.<\/p>\n<h3>Sonu\u00e7<\/h3>\n<p>Bu \"builder'\u0131n g\u00fcnl\u00fc\u011f\u00fc\" boyunca, AWS Makine \u00d6\u011frenimi d\u00fcnyas\u0131na s\u0131f\u0131rdan ad\u0131m atman\u0131n temel ad\u0131mlar\u0131n\u0131 ele ald\u0131k. Temel kavramlar\u0131 anlamaktan, AWS hesab\u0131n\u0131z\u0131 g\u00fcvenli bir \u015fekilde kurmaya, veri haz\u0131rl\u0131\u011f\u0131ndan SageMaker ile model geli\u015ftirmeye ve da\u011f\u0131tmaya kadar geni\u015f bir yelpazeyi kapsad\u0131k. Unutmay\u0131n, makine \u00f6\u011frenimi s\u00fcrekli \u00f6\u011frenmeyi gerektiren dinamik bir aland\u0131r. Bu makale size sa\u011flam bir ba\u015flang\u0131\u00e7 noktas\u0131 sunarken, ger\u00e7ek ustal\u0131k pratik yaparak, deneyler yaparak ve AWS'in s\u00fcrekli geli\u015fen hizmetlerini ke\u015ffederek gelecektir. \u015eimdi s\u0131ra sizde: kollar\u0131 s\u0131vay\u0131n ve kendi ML projelerinizi hayata ge\u00e7irmeye ba\u015flay\u0131n!<\/p>\n<h3>SSS (S\u0131k Sorulan Sorular)<\/h3>\n<h4>AWS ML'e ba\u015flamak i\u00e7in ne kadar kodlama bilgisi gerekli?<\/h4>\n<p>Temel Python bilgisi ve veri yap\u0131lar\u0131na a\u015final\u0131k genellikle yeterlidir. SageMaker SDK ve pop\u00fcler ML k\u00fct\u00fcphaneleri (Scikit-learn, Pandas) ile \u00e7al\u0131\u015fmak i\u00e7in bu temel gereklidir. Ancak Autopilot gibi hizmetler, kodlama ihtiyac\u0131n\u0131 minimuma indirir.<\/p>\n<h4>Maliyetleri nas\u0131l kontrol edebilirim?<\/h4>\n<p>AWS B\u00fct\u00e7eler olu\u015fturun, kullanmad\u0131\u011f\u0131n\u0131z kaynaklar\u0131 (notebook instance'lar\u0131, endpoint'ler) kapat\u0131n, instance tiplerini ihtiya\u00e7lar\u0131n\u0131za g\u00f6re optimize edin ve \u00fccretsiz katman avantajlar\u0131n\u0131 kullan\u0131n. Ayr\u0131ca, maliyetleri izlemek i\u00e7in AWS Cost Explorer'\u0131 d\u00fczenli olarak kontrol edin.<\/p>\n<h4>Hangi programlama dilini \u00f6\u011frenmeliyim?<\/h4>\n<p>Makine \u00f6\u011frenimi alan\u0131nda Python, a\u00e7\u0131k ara en pop\u00fcler dildir. Pandas, NumPy, Scikit-learn, TensorFlow ve PyTorch gibi g\u00fc\u00e7l\u00fc k\u00fct\u00fcphaneler sayesinde Python, ML projeleri i\u00e7in vazge\u00e7ilmezdir.<\/p>\n<h4>S\u0131f\u0131rdan ba\u015flamak ne kadar s\u00fcrer?<\/h4>\n<p>Bu tamamen sizin \u00f6\u011frenme h\u0131z\u0131n\u0131za ve ay\u0131rabilece\u011finiz zamana ba\u011fl\u0131d\u0131r. Temel kavramlar\u0131 anlamak ve ilk basit modelinizi AWS'te e\u011fitip da\u011f\u0131tmak birka\u00e7 hafta s\u00fcrebilir. Ancak ger\u00e7ek bir ML m\u00fchendisi veya veri bilimcisi olmak, aylar hatta y\u0131llar s\u00fcren s\u00fcrekli \u00f6\u011frenme ve pratik gerektirir.<\/p>\n","protected":false},"excerpt":{"rendered":"Makine \u00f6\u011frenimi ser\u00fcvenine \u00e7\u0131kmadan \u00f6nce, temel ta\u015flar\u0131 do\u011fru bir \u015fekilde yerle\u015ftirmek b\u00fcy\u00fck \u00f6nem ta\u015f\u0131r. Bu b\u00f6l\u00fcmde, makine \u00f6\u011freniminin n&#8230;","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"csco_page_header_type":"","csco_page_load_nextpost":"","csco_page_subscribe_form":"","csco_page_contact_form":"","footnotes":""},"categories":[1406],"tags":[],"class_list":{"0":"post-36571","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-aws","7":"cs-entry","8":"cs-video-wrap"},"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v20.5 (Yoast SEO v25.3.1) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Bir Builder&#039;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k - Kodlar\u0131n Gizemli D\u00fcnyas\u0131<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\" \/>\n<meta property=\"og:locale\" content=\"tr_TR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Bir Builder&#039;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k\" \/>\n<meta property=\"og:description\" content=\"Makine \u00f6\u011frenimi ser\u00fcvenine \u00e7\u0131kmadan \u00f6nce, temel ta\u015flar\u0131 do\u011fru bir \u015fekilde yerle\u015ftirmek b\u00fcy\u00fck \u00f6nem ta\u015f\u0131r. Bu b\u00f6l\u00fcmde, makine \u00f6\u011freniminin n...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\" \/>\n<meta property=\"og:site_name\" content=\"Kodlar\u0131n Gizemli D\u00fcnyas\u0131\" \/>\n<meta property=\"article:published_time\" content=\"2025-12-19T23:00:36+00:00\" \/>\n<meta name=\"author\" content=\"Fatih Soysal\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Yazan:\" \/>\n\t<meta name=\"twitter:data1\" content=\"Fatih Soysal\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tahmini okuma s\u00fcresi\" \/>\n\t<meta name=\"twitter:data2\" content=\"14 dakika\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\"},\"author\":{\"name\":\"Fatih Soysal\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"headline\":\"Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k\",\"datePublished\":\"2025-12-19T23:00:36+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\"},\"wordCount\":2571,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"articleSection\":[\"AWS\"],\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#respond\"]}],\"copyrightYear\":\"2025\",\"copyrightHolder\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\",\"name\":\"Bir Builder'\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k - Kodlar\u0131n Gizemli D\u00fcnyas\u0131\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\"},\"datePublished\":\"2025-12-19T23:00:36+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#breadcrumb\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Anasayfa\",\"item\":\"https:\/\/fatihsoysal.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/\",\"name\":\"Fatihsoysal.com\",\"description\":\"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim\",\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"tr\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\",\"name\":\"Fatih Soysal\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"tr\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"contentUrl\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"width\":512,\"height\":512,\"caption\":\"Fatih Soysal\"},\"logo\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\"},\"description\":\"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Bir Builder'\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k - Kodlar\u0131n Gizemli D\u00fcnyas\u0131","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/","og_locale":"tr_TR","og_type":"article","og_title":"Bir Builder'\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k","og_description":"Makine \u00f6\u011frenimi ser\u00fcvenine \u00e7\u0131kmadan \u00f6nce, temel ta\u015flar\u0131 do\u011fru bir \u015fekilde yerle\u015ftirmek b\u00fcy\u00fck \u00f6nem ta\u015f\u0131r. Bu b\u00f6l\u00fcmde, makine \u00f6\u011freniminin n...","og_url":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/","og_site_name":"Kodlar\u0131n Gizemli D\u00fcnyas\u0131","article_published_time":"2025-12-19T23:00:36+00:00","author":"Fatih Soysal","twitter_card":"summary_large_image","twitter_misc":{"Yazan:":"Fatih Soysal","Tahmini okuma s\u00fcresi":"14 dakika"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#article","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/"},"author":{"name":"Fatih Soysal","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"headline":"Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k","datePublished":"2025-12-19T23:00:36+00:00","mainEntityOfPage":{"@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/"},"wordCount":2571,"commentCount":0,"publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"articleSection":["AWS"],"inLanguage":"tr","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#respond"]}],"copyrightYear":"2025","copyrightHolder":{"@id":"https:\/\/fatihsoysal.com\/blog\/#organization"}},{"@type":"WebPage","@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/","url":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/","name":"Bir Builder'\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k - Kodlar\u0131n Gizemli D\u00fcnyas\u0131","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/#website"},"datePublished":"2025-12-19T23:00:36+00:00","breadcrumb":{"@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#breadcrumb"},"inLanguage":"tr","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fatihsoysal.com\/blog\/bir-builderin-gunlugu-aws-makine-ogrenimine-sifirdan-hazirlik\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Anasayfa","item":"https:\/\/fatihsoysal.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Bir Builder&#8217;\u0131n G\u00fcnl\u00fc\u011f\u00fc: AWS Makine \u00d6\u011frenimine S\u0131f\u0131rdan Haz\u0131rl\u0131k"}]},{"@type":"WebSite","@id":"https:\/\/fatihsoysal.com\/blog\/#website","url":"https:\/\/fatihsoysal.com\/blog\/","name":"Fatihsoysal.com","description":"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim","publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"tr"},{"@type":["Person","Organization"],"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1","name":"Fatih Soysal","image":{"@type":"ImageObject","inLanguage":"tr","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","contentUrl":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","width":512,"height":512,"caption":"Fatih Soysal"},"logo":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/"},"description":"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.","url":"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/"}]}},"yoast_meta":{"yoast_wpseo_title":"","yoast_wpseo_metadesc":"","yoast_wpseo_canonical":""},"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/36571","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/comments?post=36571"}],"version-history":[{"count":0,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/36571\/revisions"}],"wp:attachment":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/media?parent=36571"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/categories?post=36571"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/tags?post=36571"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}