Type Bangla using plain lowercase English letters. A bundled dictionary ranks the suggestions, so ambiguous spellings such as acho resolve to the word you meant. It runs entirely on your device, with no internet permission, no ads, and no data collection.

• Lowercase phonetic typing with dictionary-ranked suggestions
• Quick switch to English QWERTY, numbers and symbols page, Bangla digits, and emoji
• Works offline: your typing never leaves the device
• Native Kotlin input method service, no third-party tracking code

A minimalistic Android application demonstrating local, high-speed semantic search without cloud dependencies. Integrates the MiniLM-L6-V2 sentence embedding model, TensorFlow Lite, ObjectBox vector database, and Apache Lucene KNN to deliver context-aware retrieval directly on the device.

• On-device text embeddings via TensorFlow Lite & Bert vocabulary tokenizer
• Embedded vector store leveraging ObjectBox HNSW & Lucene KNN indexing
• 100% private execution with sub-30ms similarity query latency
• Clean MVVM architecture with Kotlin Coroutines and Flow

Mobile interactive testbed built on the jSciPy scientific library. Empowers mobile engineers and biomedical researchers to design digital filters (Butterworth, Chebyshev, Bessel), compute real-time FFT frequency spectrums, analyze Power Spectral Density (PSD), and detect waveform peaks.

• Zero-dependency native JVM/Android scientific mathematics engine
• Live frequency response graphing for low-pass, high-pass, and bandpass filters
• High-throughput peak detection algorithms for ECG/PPG biomedical waveforms
• Runge-Kutta (RK4) differential equation solver and numerical calculus