INFERNIX ENGINEERING JOURNAL 2026

Engineering Journal.

Notes from the workshop — how the systems on this site were actually built: architecture decisions, measured experiments, trade-offs that didn't fit in a README, and the lessons that survived contact with the benchmarks. Written by Pavan Ganesh Indupudi (Infernix).

RAG ML Systems

Inside VOICE-RAG: From Speech to Grounded Answers

A RAG system that knows when not to answer is worth more than one that always does. How calibrated per-language relevance floors, a grounding guard, and a 56 ms retrieval core came together — and where the borderline zone still hurts.

2026-10-19 9 min read
ML Systems DRAFT

Debugging Browser-Based ML Inference: Lessons from ONNX Runtime Web

Running MobileViT inside a browser extension means WebGPU when you can get it, WASM when you can't, and failure modes that only exist at the intersection. An honest draft — the debugging war stories are TODOs until the notes are verified.

2026-10-19 6 min read