<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Narcis Miclaus</title><description>Sono un data engineer in Italia, sto studiando per diventare consulente finanziario. Quello che mi piace di entrambi: costruire soluzioni, risolvere problemi, scomporre obiettivi grandi e complessi in piccoli passi chiari. Ne scrivo ogni giorno. Resta un po&apos;: magari impari qualcosa. Magari la imparo anch&apos;io.</description><link>https://narcismiclaus.com/it/</link><item><title>Capstone: cosa sai adesso, dove andare poi</title><link>https://narcismiclaus.com/it/programming/python/60-capstone/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/60-capstone/</guid><description>Uno sguardo indietro alle 60 lezioni, uno sguardo avanti a dove sta andando Python, e le risorse che ti portano da intermedio a esperto.</description><pubDate>Fri, 19 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>course-summary</category><category>next-steps</category></item><item><title>Un health check da 30 minuti su un cluster Spark che non hai mai visto</title><link>https://narcismiclaus.com/it/programming/pyspark/60-spark-cluster-health-check/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/60-spark-cluster-health-check/</guid><description>La checklist di chiusura: ti consegnano il portatile, hai tempo fino alle 17 per capire cosa non va.</description><pubDate>Thu, 18 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>dba</category><category>health-check</category><category>course-summary</category></item><item><title>AI vs ML nel 2026: quando chiamare un LLM, quando allenare</title><link>https://narcismiclaus.com/it/programming/python/59-ai-vs-ml/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/59-ai-vs-ml/</guid><description>La decisione che cinque anni fa non esisteva: usare un modello hosted, fare fine-tuning di uno open, o allenarne uno tuo?</description><pubDate>Tue, 16 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>ai</category><category>llm</category><category>machine-learning</category><category>deployment</category></item><item><title>Adaptive Query Execution: la killer feature di Spark 3.x</title><link>https://narcismiclaus.com/it/programming/pyspark/59-adaptive-query-execution/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/59-adaptive-query-execution/</guid><description>Dynamic partition coalescing, gestione dello skew a runtime e switch della strategia di join: le config da conoscere e i casi in cui AQE non può aiutarti.</description><pubDate>Mon, 15 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>aqe</category><category>optimization</category><category>performance</category></item><item><title>Modelli pre-addestrati + transfer learning + Hugging Face</title><link>https://narcismiclaus.com/it/programming/python/58-transfer-learning/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/58-transfer-learning/</guid><description>Il percorso realistico da zero a un modello di deep learning funzionante nel 2026: parti da uno pre-addestrato e fanne il fine-tuning sui tuoi dati.</description><pubDate>Fri, 12 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>deep-learning</category><category>transfer-learning</category><category>huggingface</category><category>fine-tuning</category></item><item><title>Debug di job Spark lenti: la checklist da 30 minuti</title><link>https://narcismiclaus.com/it/programming/pyspark/58-debugging-slow-jobs/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/58-debugging-slow-jobs/</guid><description>Il loop sistematico per capire cosa non va in un job lento: leggi la UI, trova lo stage lento, guarda lo skew dei task, GC, volume di shuffle, in quest&apos;ordine.</description><pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>debugging</category><category>performance</category><category>production</category></item><item><title>Il training loop, in codice</title><link>https://narcismiclaus.com/it/programming/python/57-training-loop/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/57-training-loop/</guid><description>Le cinque righe che trasformano una rete inizializzata a caso in un modello allenato, e la contabilita&apos; che le rende production-grade.</description><pubDate>Tue, 09 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>pytorch</category><category>training-loop</category><category>deep-learning</category></item><item><title>Memory tuning: executor memory, overhead, diagnostica degli OOM</title><link>https://narcismiclaus.com/it/programming/pyspark/57-memory-tuning/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/57-memory-tuning/</guid><description>I quattro config che davvero contano, cosa significa spill, come si legge uno stack trace di OOM, e la regola per dimensionare gli executor.</description><pubDate>Mon, 08 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>memory</category><category>tuning</category><category>production</category></item><item><title>PyTorch: il default moderno</title><link>https://narcismiclaus.com/it/programming/python/56-pytorch/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/56-pytorch/</guid><description>Tensor, autograd, il modulo nn, e quel feeling pythonico che ha fatto vincere PyTorch.</description><pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>pytorch</category><category>deep-learning</category><category>tensors</category></item><item><title>Leggere gli execution plan: .explain(True), dal parsed al physical</title><link>https://narcismiclaus.com/it/programming/pyspark/56-execution-plans/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/56-execution-plans/</guid><description>Come leggere ogni riga dell&apos;output di .explain(), gli operatori che contano, e i passi dell&apos;optimizer che li producono.</description><pubDate>Thu, 04 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>explain</category><category>execution-plan</category><category>catalyst</category></item><item><title>Neural network spiegate semplice</title><link>https://narcismiclaus.com/it/programming/python/55-neural-networks-plain/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/55-neural-networks-plain/</guid><description>Cos&apos;è davvero una neural network, perché la backpropagation funziona, e dove il deep learning batte sul serio il machine learning classico.</description><pubDate>Tue, 02 Jun 2026 00:00:00 GMT</pubDate><category>python</category><category>deep-learning</category><category>neural-networks</category><category>fundamentals</category></item><item><title>La Spark UI: lo strumento più importante che imparerai</title><link>https://narcismiclaus.com/it/programming/pyspark/55-spark-ui/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/55-spark-ui/</guid><description>Un giro guidato di ogni tab (Jobs, Stages, Tasks, SQL, Storage, Executors) e cosa ti dice ognuno quando qualcosa va storto.</description><pubDate>Mon, 01 Jun 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>ui</category><category>debugging</category><category>production</category></item><item><title>ML project: un problema di classificazione, end to end</title><link>https://narcismiclaus.com/it/programming/python/54-ml-project/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/54-ml-project/</guid><description>Da CSV grezzo a modello in produzione: le lezioni del Modulo 9 rese tangibili.</description><pubDate>Fri, 29 May 2026 00:00:00 GMT</pubDate><category>python</category><category>machine-learning</category><category>project</category><category>end-to-end</category><category>classification</category></item><item><title>Output mode e sink idempotenti: foreachBatch e il pattern di upsert</title><link>https://narcismiclaus.com/it/programming/pyspark/54-output-modes-and-sinks/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/54-output-modes-and-sinks/</guid><description>Append vs update vs complete, i sink che Spark fornisce, e l&apos;escape hatch foreachBatch per tutto il resto.</description><pubDate>Thu, 28 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>streaming</category><category>sinks</category><category>idempotent</category><category>foreach-batch</category></item><item><title>Hyperparameter tuning: grid, random, bayesian, optuna</title><link>https://narcismiclaus.com/it/programming/python/53-hyperparameter-tuning/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/53-hyperparameter-tuning/</guid><description>Le quattro strategie di ricerca, quando ognuna ha senso, e perche&apos; optuna e&apos; il default del 2026.</description><pubDate>Tue, 26 May 2026 00:00:00 GMT</pubDate><category>python</category><category>machine-learning</category><category>hyperparameters</category><category>optuna</category><category>gridsearch</category></item><item><title>Operazioni stateful: aggregazioni, sessioni e lo state store</title><link>https://narcismiclaus.com/it/programming/pyspark/53-stateful-operations/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/53-stateful-operations/</guid><description>Dove Spark Streaming tiene lo stato tra micro-batch, i pattern stateful standard, e quando scendere a mapGroupsWithState.</description><pubDate>Mon, 25 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>streaming</category><category>state</category><category>sessionization</category></item><item><title>Modelli lineari e regolarizzati: quando il semplice vince</title><link>https://narcismiclaus.com/it/programming/python/52-linear-models/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/52-linear-models/</guid><description>Perche&apos; i modelli lineari sono ancora la risposta giusta sorprendentemente spesso, e i trucchi di regularization che li rendono pronti per la produzione.</description><pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate><category>python</category><category>machine-learning</category><category>linear-regression</category><category>logistic-regression</category><category>regularization</category></item><item><title>Watermark ed event time: la parte che quasi tutti i principianti sbagliano</title><link>https://narcismiclaus.com/it/programming/pyspark/52-watermarks-event-time/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/52-watermarks-event-time/</guid><description>Perché l&apos;event time conta più del processing time, cosa fa davvero un watermark, e l&apos;esempio guidato con timestamp concreti.</description><pubDate>Thu, 21 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>streaming</category><category>watermarks</category><category>event-time</category></item><item><title>Modelli ad albero: random forest, XGBoost, LightGBM</title><link>https://narcismiclaus.com/it/programming/python/51-tree-models/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/51-tree-models/</guid><description>Perché gli alberi dominano il ML tabulare, le differenze tra le tre grandi librerie, e gli iperparametri che contano.</description><pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate><category>python</category><category>machine-learning</category><category>random-forest</category><category>xgboost</category><category>lightgbm</category><category>gradient-boosting</category></item><item><title>Kafka source: l&apos;ingest di produzione più comune</title><link>https://narcismiclaus.com/it/programming/pyspark/51-kafka-source/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/51-kafka-source/</guid><description>Come Spark legge da Kafka, la semantica degli offset, e la questione at-least-once vs exactly-once.</description><pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>kafka</category><category>streaming</category><category>structured-streaming</category></item><item><title>Feature engineering: la parte che conta di più</title><link>https://narcismiclaus.com/it/programming/python/50-feature-engineering/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/50-feature-engineering/</guid><description>Le trasformazioni che trasformano i dati grezzi in carburante per il modello, e quelle che silenziosamente fanno trapelare informazioni dal futuro.</description><pubDate>Fri, 15 May 2026 00:00:00 GMT</pubDate><category>python</category><category>machine-learning</category><category>features</category><category>preprocessing</category></item><item><title>Structured Streaming: le basi di readStream, writeStream, trigger</title><link>https://narcismiclaus.com/it/programming/pyspark/50-structured-streaming-basics/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/50-structured-streaming-basics/</guid><description>Gli entry point per lo streaming, la semantica dei trigger, e il checkpoint da cui dipende tutto.</description><pubDate>Thu, 14 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>streaming</category><category>structured-streaming</category><category>dataframe</category></item><item><title>scikit-learn: il tour della libreria standard di ML</title><link>https://narcismiclaus.com/it/programming/python/49-scikit-learn/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/49-scikit-learn/</guid><description>Il pattern fit/predict che sta dietro a tutto, le categorie di modelli, e le pipeline che rendono il codice riproducibile.</description><pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate><category>python</category><category>scikit-learn</category><category>machine-learning</category><category>ml</category></item><item><title>Perché lo streaming, e cosa significa davvero &apos;streaming&apos; in Spark</title><link>https://narcismiclaus.com/it/programming/pyspark/49-why-streaming/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/49-why-streaming/</guid><description>Dati bounded vs unbounded, batch e streaming come continuum, e perché i DStreams sono deprecati a favore di Structured Streaming.</description><pubDate>Mon, 11 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>streaming</category><category>structured-streaming</category><category>fundamentals</category></item><item><title>Progetto numerico: un&apos;analisi vera</title><link>https://narcismiclaus.com/it/programming/python/48-numerical-project/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/48-numerical-project/</guid><description>Prendi un dataset pubblico ed esegui un&apos;analisi numerica completa: statistiche descrittive, fit, test di ipotesi, plot.</description><pubDate>Fri, 08 May 2026 00:00:00 GMT</pubDate><category>python</category><category>numpy</category><category>scipy</category><category>project</category><category>analysis</category></item><item><title>Schema evolution: quando le colonne ti cambiano sotto</title><link>https://narcismiclaus.com/it/programming/pyspark/48-schema-evolution/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/48-schema-evolution/</guid><description>Perché i formati schema-on-read gestiscono male il cambiamento, perché Avro+registry lo gestisce bene, e il modo Delta/Iceberg di avere entrambe le cose.</description><pubDate>Thu, 07 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>schema</category><category>parquet</category><category>avro</category><category>evolution</category></item><item><title>Jupyter, notebook, e quando lasciarli</title><link>https://narcismiclaus.com/it/programming/python/47-jupyter-and-notebooks/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/47-jupyter-and-notebooks/</guid><description>Perche&apos; i notebook danno dipendenza, dove brillano e il momento in cui dovresti fermarti e scrivere uno script vero.</description><pubDate>Tue, 05 May 2026 00:00:00 GMT</pubDate><category>python</category><category>jupyter</category><category>notebooks</category><category>vscode</category><category>kernel</category></item><item><title>Cloud storage: S3, GCS, Azure Blob, cosa cambia</title><link>https://narcismiclaus.com/it/programming/pyspark/47-cloud-storage/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/47-cloud-storage/</guid><description>Le note in piccolo sulla consistency, il problema del rename, e perche&apos; esistono i committer direct-write.</description><pubDate>Mon, 04 May 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>s3</category><category>cloud</category><category>storage</category><category>hadoop</category></item><item><title>Capstone: progettare un&apos;architettura completa per un&apos;azienda fittizia a tre scale</title><link>https://narcismiclaus.com/it/programming/architecture/80-capstone/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/80-capstone/</guid><description>Ottanta lezioni di system architecture, condensate in un singolo esercizio di design. La stessa azienda SaaS fittizia, tre scale, tre architetture, e un tour guidato di cosa cambia e perché. La lezione di chiusura del corso.</description><pubDate>Fri, 01 May 2026 00:00:00 GMT</pubDate><category>architecture</category><category>capstone</category><category>course-summary</category></item><item><title>Le funzionalità di Python che ho imparato troppo tardi</title><link>https://narcismiclaus.com/it/programming/python/features-learned-too-late/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/features-learned-too-late/</guid><description>Match statement, il walrus operator, il debug con f-string, le dataclass e altre funzionalità di Python che mi avrebbero risparmiato ore se le avessi conosciute prima.</description><pubDate>Fri, 01 May 2026 00:00:00 GMT</pubDate><category>python</category><category>tips</category><category>modern-python</category></item><item><title>Scrivere su JDBC: parallelismo, batch, idempotenza</title><link>https://narcismiclaus.com/it/programming/pyspark/46-writing-jdbc/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/46-writing-jdbc/</guid><description>Come riscrivere l&apos;output di Spark in un database relazionale senza schiacciarlo, rompere transazioni o perdere dati al retry.</description><pubDate>Thu, 30 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>jdbc</category><category>write</category><category>transactions</category></item><item><title>SciPy: la cassetta degli attrezzi che quasi tutti dimenticano</title><link>https://narcismiclaus.com/it/programming/python/45-scipy/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/45-scipy/</guid><description>Statistica, ottimizzazione, signal processing, matrici sparse: la libreria standard del Python scientifico.</description><pubDate>Tue, 28 Apr 2026 00:00:00 GMT</pubDate><category>python</category><category>scipy</category><category>statistics</category><category>optimization</category></item><item><title>Architettura di una ML platform: feature store, model registry, serving</title><link>https://narcismiclaus.com/it/programming/architecture/79-ml-platform/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/79-ml-platform/</guid><description>I cinque layer su cui una ML platform moderna si è standardizzata, il problema del train-serve skew per cui è stato inventato il feature store, e il calcolo build-versus-buy per ogni layer nel 2026.</description><pubDate>Mon, 27 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>ml-platform</category><category>feature-store</category><category>mlflow</category><category>serving</category></item><item><title>Leggere da JDBC: estrarre da Postgres, MySQL, SQL Server</title><link>https://narcismiclaus.com/it/programming/pyspark/45-reading-jdbc/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/45-reading-jdbc/</guid><description>Il connettore source JDBC, il trucco di partitionColumn, e perché una lettura ingenua manda al tappeto il database sorgente.</description><pubDate>Mon, 27 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>jdbc</category><category>postgres</category><category>mysql</category><category>parallel-read</category></item><item><title>Privacy e compliance: GDPR, CCPA, data residency</title><link>https://narcismiclaus.com/it/programming/architecture/78-privacy-compliance/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/78-privacy-compliance/</guid><description>Le normative sulla privacy come driver architetturali. Diritto all&apos;oblio, data residency, customer-managed keys, e l&apos;infrastruttura di consent e audit che i framework di compliance richiedono.</description><pubDate>Fri, 24 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>gdpr</category><category>ccpa</category><category>privacy</category><category>compliance</category><category>residency</category></item><item><title>Plotting: matplotlib, seaborn, plotly, scegliere il tuo</title><link>https://narcismiclaus.com/it/programming/python/44-plotting/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/44-plotting/</guid><description>Tre librerie di plotting, tre filosofie, e quale prendere in base al pubblico.</description><pubDate>Fri, 24 Apr 2026 00:00:00 GMT</pubDate><category>python</category><category>matplotlib</category><category>seaborn</category><category>plotly</category><category>visualization</category></item><item><title>ORC, Avro, Delta: le alternative e quando ognuna vince</title><link>https://narcismiclaus.com/it/programming/pyspark/44-orc-avro-delta/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/44-orc-avro-delta/</guid><description>Tre famiglie di formati che non sono Parquet, quando ognuna è la scelta giusta, e perché Delta sta silenziosamente prendendo il sopravvento.</description><pubDate>Thu, 23 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>orc</category><category>avro</category><category>delta</category><category>file-format</category></item><item><title>Security architecture: least privilege, defense in depth</title><link>https://narcismiclaus.com/it/programming/architecture/77-security-architecture/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/77-security-architecture/</guid><description>I principi di sicurezza di cui ogni sistema ha bisogno come architettura portante. Least privilege, defense in depth, zero trust, e i controlli IAM e di rete che trasformano i principi in realtà.</description><pubDate>Wed, 22 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>security</category><category>iam</category><category>least-privilege</category></item><item><title>NumPy: array, broadcasting, le fondamenta del Python scientifico</title><link>https://narcismiclaus.com/it/programming/python/43-numpy/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/43-numpy/</guid><description>Cos&apos;è un ndarray, perché il broadcasting cambia il modo in cui scrivi i loop, e il piccolo insieme di funzioni che copre la maggior parte dei casi.</description><pubDate>Tue, 21 Apr 2026 00:00:00 GMT</pubDate><category>python</category><category>numpy</category><category>arrays</category><category>broadcasting</category><category>vectorization</category></item><item><title>Disaster recovery: RTO, RPO, il drill</title><link>https://narcismiclaus.com/it/programming/architecture/76-disaster-recovery/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/76-disaster-recovery/</guid><description>Cosa significa davvero in pratica il disaster recovery. Le quattro DR tier, RTO e RPO come manopole di design, e la disciplina del drill che dimostra che il piano funziona.</description><pubDate>Mon, 20 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>disaster-recovery</category><category>rto</category><category>rpo</category><category>backups</category></item><item><title>Parquet: perché è il default per un buon motivo</title><link>https://narcismiclaus.com/it/programming/pyspark/43-parquet/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/43-parquet/</guid><description>Lo storage columnar spiegato, codec di compressione, predicate pushdown e la struttura a row-group che rende veloci le letture selettive.</description><pubDate>Mon, 20 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>parquet</category><category>file-format</category><category>columnar</category></item><item><title>Deployment multi-region: active-active, active-passive, follow-the-sun</title><link>https://narcismiclaus.com/it/programming/architecture/75-multi-region/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/75-multi-region/</guid><description>Perché i team vanno in multi-region (latenza, DR, compliance, capacità), le tre forme di deployment, i problemi difficili (replica, conflitti, costo), e quando non vale la pena.</description><pubDate>Fri, 17 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>multi-region</category><category>geography</category><category>latency</category></item><item><title>Progetto di data engineering: costruisci una pipeline vera</title><link>https://narcismiclaus.com/it/programming/python/42-data-engineering-project/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/42-data-engineering-project/</guid><description>Dalla sorgente alla destinazione, con monitoring, idempotenza e una schedulazione. Le lezioni del Modulo 7 rese tangibili.</description><pubDate>Fri, 17 Apr 2026 00:00:00 GMT</pubDate><category>python</category><category>data-engineering</category><category>pipeline</category><category>project</category><category>end-to-end</category></item><item><title>Tungsten: code generation e layout di memoria colonnare</title><link>https://narcismiclaus.com/it/programming/pyspark/42-tungsten/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/42-tungsten/</guid><description>Come Spark fonde le operazioni in codice compilato, il formato colonnare off-heap, e perché DataFrame Spark è veloce.</description><pubDate>Thu, 16 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>tungsten</category><category>performance</category><category>internals</category></item><item><title>Health check di SQL Server in 30 minuti: lo script che esegui davvero</title><link>https://narcismiclaus.com/it/programming/sql-server/40-health-check/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/sql-server/40-health-check/</guid><description>La checklist ordinata che eseguo su qualsiasi SQL Server che non ho mai visto prima. Stampabile, copiabile, e la lezione di chiusura del corso.</description><pubDate>Thu, 16 Apr 2026 00:00:00 GMT</pubDate><category>sql-server</category><category>dba</category><category>health-check</category><category>course-summary</category></item><item><title>Perché scrivo di finanza personale</title><link>https://narcismiclaus.com/it/finance/why-im-writing-about-money/</link><guid isPermaLink="true">https://narcismiclaus.com/it/finance/why-im-writing-about-money/</guid><description>Perché un data engineer apre una sezione di finanza sul suo sito, e che genere di scrittura aspettarsi qui.</description><pubDate>Thu, 16 Apr 2026 00:00:00 GMT</pubDate><category>intro</category><category>personal-finance</category></item><item><title>Event-driven architecture: saga, choreography, orchestration</title><link>https://narcismiclaus.com/it/programming/architecture/74-event-driven-architecture/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/74-event-driven-architecture/</guid><description>Servizi che si parlano emettendo eventi, la scelta tra choreography e orchestration, il pattern saga, e il toolset 2026 (Temporal, Step Functions, Camunda, Argo).</description><pubDate>Wed, 15 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>event-driven</category><category>saga</category><category>choreography</category><category>orchestration</category></item><item><title>Orchestrazione: Airflow, Prefect, Dagster, il panorama del 2026</title><link>https://narcismiclaus.com/it/programming/python/41-orchestration/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/python/41-orchestration/</guid><description>Quando cron non basta, cosa fa davvero un orchestrator, e i tre contendenti per le tue pipeline di dati.</description><pubDate>Tue, 14 Apr 2026 00:00:00 GMT</pubDate><category>python</category><category>airflow</category><category>prefect</category><category>dagster</category><category>orchestration</category></item><item><title>Microservices: quando, quando no, il ritorno del monolith</title><link>https://narcismiclaus.com/it/programming/architecture/73-microservices/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/architecture/73-microservices/</guid><description>Il boom dei microservices del 2015-2020, il pushback dal 2021 in poi, e il modular monolith come via di mezzo. La legge di Conway, la tassa dei sistemi distribuiti, e come scegliere in base alla dimensione del team e al profilo di scaling.</description><pubDate>Mon, 13 Apr 2026 00:00:00 GMT</pubDate><category>architecture</category><category>microservices</category><category>monolith</category></item><item><title>Catalyst: il cervello dietro ogni DataFrame</title><link>https://narcismiclaus.com/it/programming/pyspark/41-catalyst-optimizer/</link><guid isPermaLink="true">https://narcismiclaus.com/it/programming/pyspark/41-catalyst-optimizer/</guid><description>Come Spark trasforma il tuo codice in un query plan, le quattro fasi di ottimizzazione, e come leggere .explain(True).</description><pubDate>Mon, 13 Apr 2026 00:00:00 GMT</pubDate><category>pyspark</category><category>spark</category><category>catalyst</category><category>optimizer</category><category>explain</category></item></channel></rss>