About — Shubham Singh Bisht

Background & Focus

I have always been drawn to the tactile nature of mobile and embedded devices—the fact that code running on silicon translates directly to someone's fingers reacting on a slab of glass. Over the years, that curiosity pulled me deeper into the stack: moving from building modular consumer apps to investigating ART runtime behavior, optimizing Linux socket multiplexing, and tuning neural co-processors.

In open-source, I build and experiment as @AndroidLord on GitHub, maintaining over 60 repositories exploring Android architecture patterns, Kotlin coroutine mechanics, scrapers, and backend APIs.

Experience & Journey

Teachmint

Android Engineer · Systems & Large Screens

Bengaluru

Focused on core Android client software and custom system layers for large-screen interactive touch displays and dedicated hardware appliances:

High-Performance Canvas Rendering: Engineered low-latency drawing and multi-touch pipelines for large-format 4K displays. Decoupled high-frequency touch ingestion from UI layout passes, applied primitive ring buffers to eliminate GC pauses, and achieved sub-30ms touch-to-pixel latency at 60 fps.
Cross-App Identity Architecture: Implemented a unified Single Sign-On (SSO) layer coordinating multiple standalone applications on shared devices via Android's native AccountManager and signature-level IPC, eliminating redundant logins and preventing concurrent token-refresh race conditions.
Edge Intelligence Prototyping: Explored on-device small language model (SLM) inference and local vector retrieval on embedded ARM silicon, utilizing dedicated NPU acceleration to maintain 60 fps UI responsiveness during active local generation.

Pratilipi

Android Engineer · Mobile Platform

Bengaluru

Contributed to high-scale consumer Android applications serving millions of readers across diverse device tiers:

Architected and modernized product features using clean MVVM architecture, Jetpack Compose, Coroutines, and reactive StateFlow streams.
Profiled and reduced cold application startup latency, eliminated memory leaks on resource-constrained devices, and optimized offline reading caches for spotty network conditions.

Technical Toolkit

Languages

Kotlin, Java, C/C++ (Android NDK), Python, SQL

Android & Framework

Jetpack Compose, Android SurfaceView, Coroutines & Flow, AccountManager IPC, Room, Retrofit, Dagger Hilt

Systems & Profiling

Android Runtime (ART), Perfetto, Systrace, ARM NEON SIMD, epoll sockets, Memory & GC Profiling

Edge AI & ML

ONNX Runtime Mobile, RKNN NPU Toolkit, SQLite FTS5 Vector Hybrid Search, Model Quantization (INT4 / INT8)

Engineering Interests

Android framework, AOSP internals, and ART runtime mechanics
Concurrency, coroutines, and thread scheduling algorithms
Low-latency graphics rendering on large-screen & dedicated hardware
On-device AI inference, quantization, and local RAG retrieval
Pragmatic architecture that earns its abstraction layers

Working Philosophy

I believe in investigation over guesswork. When a frame drops, a memory leak surfaces, or a coroutine deadlocks, guessing at fixes introduces accidental complexity. Real engineering happens by digging into the mechanics—reading framework source code, capturing Systrace and Perfetto profiles, and understanding the runtime contract.

Every robust system is built on an honest post-mortem loop: assumption → failure → investigation → correction. I respect architectures that earn their layers, prioritize maintainability over novelty, and favor simple, verifiable invariants.

Currently Focusing On

Building: Android software & system layers for interactive flat-panel (IFP) devices
Exploring: On-device LLM/SLM inference, NPU acceleration, and local RAG pipelines
Investigating: Android runtime (ART), coroutine schedulers, and IPC boundaries

Get in Touch

I am always open to discussing Android systems, edge computing, or technical writing. Feel free to connect: