Technical articles and historical engineering writing on software and systems engineering, real-time data processing, artificial intelligence, speech and signal processing, digital forensics, cybersecurity and algorithms.
The Blog contains my technical articles on software and systems engineering, artificial intelligence, digital forensics, cybersecurity, and the history of technology.
Host-level overwrite does not prove SSD sanitization because logical blocks are decoupled from physical NAND by wear leveling, remapping and over-provisioning.
WABT exposes WebAssembly sections, imports, exports and structured control flow without forcing a PE/ELF analysis model onto WASM.
FPGA LLM inference requires algorithm/hardware co-design across quantization, memory bandwidth, pipelining, on-chip buffers and KV cache.
Concurrency in distributed data systems depends on the interaction of MVCC, isolation levels, lost updates, write skew, and replication visibility rather than on a single database setting.
Whisper converts 16 kHz audio to a log-Mel spectrogram and generates tokens autoregressively with an encoder-decoder Transformer; language, task, and timestamp information share the same token sequence.
NetMQ removes the broker requirement, not the need to define queues, backpressure, socket ownership, acknowledgement and reconnect semantics.
An essay on the distinction among data, information, knowledge, and wisdom through historical development, experience, uncertainty, and responsibility in decision-making.
Turkish syllabification can be implemented in linear time by tracking vowel and consonant transitions with a small state machine; the difficult cases arise in compounds and non-linguistic input.
This C# WebP encoder produces the VP8L bitstream directly without wrapping native libwebp; transforms, LZ77 search, and canonical Huffman coding form the core pipeline.
Variance-based color quantization uses moments over RGBA space and repeatedly splits the boxes carrying the greatest error to build a limited palette.
Processing OCR and ASR output through morphology, syntax, semantics, n-gram/TF-IDF representations, and dialogue analysis, then linking extraction and search to a multimodal timeline.
Hardware layers that determine multimedia performance, from sensors and ADCs to CPUs, SIMD, GPUs, memory systems and embedded accelerators.
Uses only this site's published content. How it works
