In this episode, Lucas and Luna explore how Android developers are using Kotlin Flow for reactive UI patterns in 2026. They break down a real example: a food delivery app that eliminated manual state refreshing by using Flow to observe live order status changes. Lucas explains the difference between StateFlow and SharedFlow, and why cold flows matter for database queries. Luna shares how iOS developers are now bringing similar patterns to Swift with AsyncSequence. The hosts also discuss how reactive UI reduces boilerplate, improves battery life by stopping unnecessary recompositions, and changes how developers think about data pipelines. If you're building mobile apps and want cleaner state management, this episode gives you a concrete pattern you can start using today.

#KotlinFlow #ReactiveUI #StateFlow #SharedFlow #AndroidDev #SwiftAsyncSequence #MobileApps #UIStateManagement #FoodDeliveryApp #LiveOrderTracking #ColdFlow #HotFlow #JetpackCompose #SwiftUI #Technology #MobileDevelopment #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore how mobile apps are using on-device AI to modulate voices in real time—changing pitch, tone, and accent during calls, recordings, and live streaming. They break down the technology behind voice conversion models that run entirely on the phone, preserving privacy while enabling new use cases from accessibility to entertainment. The episode focuses on the specific case of a 2025 startup called VoixAI that raised $12 million for a real-time voice modulation SDK, and how major messaging apps like Signal and Telegram are testing similar features. Lucas explains the latency challenges—how developers achieve sub-100-millisecond processing using neural audio codecs and transformer-based voice encoders—and Luna questions whether voice modulation could erode trust in phone calls. They also touch on regulatory implications, including the EU's proposed mandatory watermarking for AI-altered audio by 2027.

#VoiceModulation #OnDeviceAI #RealTimeAudio #VoixAI #MobileApps #NeuralAudioCodec #VoiceConversion #PrivacyPreserving #Signal #Telegram #Accessibility #DeepfakeRegulation #AIWatermarking #LowLatency #TransformerModel #TechPodcast #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode of Mobile Development with Fexingo, Lucas and Luna explore how mobile apps are using on-device AI to summarize text in real time. They break down the technical challenges—like running large language models on a phone without draining the battery—and highlight concrete examples, from news apps that condense articles into three bullet points to messaging apps that summarize group chats. They discuss the privacy advantage of keeping summarization on-device versus sending data to the cloud, and look at how Apple and Google are building these capabilities into their mobile operating systems. Lucas shares a specific benchmark: the latest Qualcomm Snapdragon chip can run a 7-billion-parameter model at 30 tokens per second, making real-time summarization feasible. The hosts also touch on the trade-offs in accuracy and the need for fine-tuning on device-specific data. Tune in for a focused, practical look at a feature that's quietly changing how we consume information on mobile.

#OnDeviceAI #TextSummarization #MobileApps #LLM #Snapdragon #Qualcomm #Apple #Google #PrivacvPreservingAI #RealTimeProcessing #iOSDevelopment #AndroidDevelopment #FexingoBusiness #BusinessPodcast #Technology #AIOnPhone #EdgeAI #SummarizationApp

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into the growing trend of mobile fitness apps using on-device AI to provide real-time coaching without cloud latency. They examine the case of the popular app FitForm, which uses a smartphone camera and on-device machine learning models to analyze exercise form in real time, offering corrective feedback on squats, push-ups, and yoga poses. The episode covers how this technology works under the hood, including pose estimation models like MoveNet and TensorFlow Lite, and discusses the privacy benefits of keeping all video processing on the device. They also touch on the competitive landscape, with Apple's Fitness+ and Google's Fitbit both exploring similar on-device features. Lucas and Luna debate whether this approach truly replaces a human coach, and explore the limitations around body type bias in training data. A natural donation segment ties the value of privacy-preserving AI to listener support that keeps the show ad-free.

#OnDeviceAI #FitnessTech #MobileApps #RealTimeCoaching #FitForm #PoseEstimation #MoveNet #TensorFlowLite #PrivacyFirst #AppleFitnessPlus #Fitbit #MachineLearning #ComputerVision #EdgeAI #TechPodcast #MobileDevelopment #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 47 of Mobile Development with Fexingo explores how mobile apps are now using on-device AI for real-time lip reading. Lucas and Luna break down the technology behind visual speech recognition, its current accuracy rates, and practical use cases like accessibility for the deaf and hard-of-hearing, silent command input in noisy environments, and privacy-preserving authentication. They discuss the neural network architecture, training data requirements, and the trade-offs between accuracy and latency. Specific examples include Google's Live Caption expansion and a startup called ReadAlong that uses lip-reading AI for language learning. The hosts also touch on ethical concerns around consent and surveillance, and how on-device processing mitigates privacy risks. As always, no ads — listener-supported at buy me a coffee dot com slash fexingo.

#MobileApps #OnDeviceAI #LipReading #VisualSpeechRecognition #Accessibility #DeafCommunity #Privacy #AI #MachineLearning #NeuralNetworks #Google #LiveCaption #ReadAlong #LanguageLearning #Technology #FexingoBusiness #BusinessPodcast #MobileDevelopment

Keep every episode free: buymeacoffee.com/fexingo

In this episode of Mobile Development with Fexingo, Lucas and Luna dive into how mobile apps are leveraging on-device AI for real-time language translation, moving beyond cloud-dependent models. They explore the technical shift from server-based translation to on-device neural networks, focusing on a concrete example: the translation feature in WhatsApp that now runs entirely on-device, reducing latency and improving privacy. The hosts discuss the model compression techniques that make this possible, including quantization and pruning, and touch on the trade-offs—like accuracy versus speed. They also look at how this impacts user experience in messaging apps, with real-time translation happening as you type, and consider the broader implications for global communication. Lucas shares that WhatsApp's on-device model is about 100 megabytes, a fraction of its earlier cloud-dependent size, while Luna questions whether smaller models sacrifice nuance. The episode ends with a reflection on how on-device AI is making translation more accessible and private, and a brief, conversational mention of listener support via buy me a coffee dot com slash fexingo.

#OnDeviceAI #MachineTranslation #RealTimeTranslation #MobileApps #NeuralNetworks #ModelCompression #Quantization #Pruning #Privacy #WhatsApp #iOS #Android #NLP #EdgeComputing #LanguageModels #Technology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 45 of Mobile Development with Fexingo dives into on-device AI-powered real-time video upscaling. Lucas and Luna unpack how Qualcomm's Snapdragon 8 Gen 4 chip and Apple's A18 Bionic enable 1080p to 4K upscaling locally, without cloud latency. They explore the neural engine architecture, the power draw improvements—down to 2.3 watts per hour—and why this matters for streaming apps, social platforms, and video editing tools. Specific benchmarks from the Samsung Galaxy S26 and iPhone 17 Pro Max are discussed, along with developer integration via Metal Performance Shaders and Qualcomm's SNPE SDK. The episode also addresses the privacy win: no video frames leave the device. A no-ads principle is briefly supported by listener contributions at buy me a coffee dot com slash fexingo.

#MobileApps #OnDeviceAI #VideoUpscaling #RealTimeProcessing #Snapdragon8Gen4 #A18Bionic #SamsungGalaxyS26 #iPhone17ProMax #MetalPerformanceShaders #QualcommSNPE #Privacy #MachineLearning #AppDevelopment #Technology #FexingoBusiness #BusinessPodcast #MobileDev #EdgeAI

Keep every episode free: buymeacoffee.com/fexingo

On this episode of Mobile Development with Fexingo, Lucas and Luna dive into the growing trend of on-device AI for real-time transcription in mobile apps. They explore why developers are moving transcription processing from the cloud to the device, focusing on privacy, speed, and offline accessibility. Lucas breaks down the technical challenges, like balancing accuracy with latency on mobile chips, and highlights real-world use cases such as live captions in video calls, voice note transcription, and accessibility tools. Luna questions whether on-device models can match cloud-based accuracy and shares how Apple's Whisper-based model and Google's Live Caption are setting the standard. The hosts also discuss the implications for app developers, including the need to manage battery life and model size. A practical episode for anyone building—or using—transcription features in their apps.

#MobileApps #OnDeviceAI #RealTimeTranscription #SpeechRecognition #AppDevelopment #iOS #Android #MachineLearning #Privacy #Accessibility #LiveCaptions #WhisperModel #GoogleLiveCaption #VoiceNotes #AppDev #FexingoBusiness #BusinessPodcast #Technology

Keep every episode free: buymeacoffee.com/fexingo

In this episode of Mobile Development with Fexingo, Lucas and Luna explore how mobile apps are using federated learning to train AI models on user data without ever sending that data to the cloud. They break down the technical architecture—how model updates are computed on-device, aggregated, and averaged—and discuss real-world implementations from Google's Gboard keyboard to Apple's QuickType and on-device Siri. The hosts also examine the trade-offs: better privacy and lower latency versus the challenges of non-i.i.d. data, device heterogeneity, and communication efficiency. They explain why federated learning is becoming a cornerstone of privacy-preserving mobile AI in 2026, especially as regulations like GDPR and China's Personal Information Protection Law tighten data transfer rules. The conversation includes a concrete example of a health app using federated learning to improve activity predictions without uploading sensitive step or location data. Lucas and Luna also touch on how the technology impacts developers—what changes in the training pipeline, how to handle model personalization, and why on-device evaluation is critical. A thoughtful, specific look at a paradigm shift in mobile machine learning.

#FederatedLearning #MobileApps #PrivacyPreservingAI #OnDeviceAI #GoogleGboard #AppleQuickType #GDPR #PIPL #MachineLearning #DataPrivacy #MobileDevelopment #Technology #FexingoBusiness #BusinessPodcast #AI #UserPrivacy #ModelTraining #EdgeAI

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore how mobile apps are moving ad targeting from the cloud to the device, using on-device AI models that analyze user behavior without sending data to servers. They break down Google's Privacy Sandbox for Android and Apple's SKAdNetwork as real-world frameworks, and discuss the trade-offs for developers: less precise targeting but better privacy compliance. The hosts also touch on how this shift affects ad revenue and user trust, with concrete examples from the current mobile ad landscape in June 2026. A practical look at a quiet revolution in mobile advertising.

#OnDeviceAI #PrivacyPreservingAds #MobileAdvertising #GooglePrivacySandbox #SKAdNetwork #Apple #Google #AdTargeting #UserPrivacy #MobileDev #iOSDevelopment #AndroidDevelopment #Technology #BusinessPodcast #FexingoBusiness #Podcast #MobileApps #AdTech

Keep every episode free: buymeacoffee.com/fexingo