Skip to content
The AppScale ArchiveWriting on Sovereign AI
21.4934° N / 86.9135° EEST. 2025 — India
300+ Essays · 27 Series
Scroll ↓

OldKnowledge,New Vessel.

Essays on on-device AI, data sovereignty, and building systems that keep knowledge where it belongs.

Enter the archive
— 00 / ThesisEvery business runs on knowledge older than its software.
01The Archive

Latest Entries

Full index — 300 essays →
02The Library
श्रीगणेशाय नमः ॥
अथ प्रथमोऽध्यायः ॥
विद्या ददाति विनयं विनयाद् याति पात्रताम् ।
पात्रत्वाद्धनमाप्नोति धनाद्धर्मं ततः सुखम् ॥
[ Coming Soon ]

We digitize centuries-old manuscripts. Then we build with the same discipline.

AppScale's roots are in a quiet project: structuring classical Sanskrit texts into faithful digital form. Extraction, structure, provenance, sovereignty — the same principles now power our client work.

न हि ज्ञानेन सदृशं पवित्रमिह विद्यते ।
“Nothing in this world purifies like knowledge.”
Bhagavad Gita · 4.38
03Capabilities

Built for Every Business

Executive AI Series · MCP Security · 日本語 · Edge AI Engineering ·
RAG in Production · Sovereign AI · Fine-Tuning · Agentic Systems ·
19+ yrs engineering·npm — react-native-edge-vector-store·Read in IN · JP · SG·AppScale LLP — DPIIT recognized
05Full Index
KV-Cache Offloading: Serving 10x More Users by Not Recomputing
ai-architecture1 min read

KV-Cache Offloading: Serving 10x More Users by Not Recomputing

Your GPU re-prefills the same 15,000-token prompt ten thousand times a day. KV-cache offloading to DRAM and NVMe turns that recompute into a cheap fetch — 10x users.

July 13, 2026Read
Feature Flag Architecture: Rollouts, Kill Switches, and Flag Debt
ai-architecture1 min read

Feature Flag Architecture: Rollouts, Kill Switches, and Flag Debt

Ship code to production without releasing it to everyone. Feature flags done right: release vs kill-switch types, sticky rollouts, failure defaults, and killing flag debt.

July 13, 2026Read
Distributed Locks: Redlock, Fencing Tokens, and Why Your Lock Doesn't Lock
ai-architecture1 min read

Distributed Locks: Redlock, Fencing Tokens, and Why Your Lock Doesn't Lock

A billing system double-charged 14 customers through a lock that worked exactly as documented. Redlock, fencing tokens, advisory locks, and when to delete the lock.

July 13, 2026Read
The CAP Theorem, PACELC, and What Real Databases Actually Choose
ai-architecture1 min read

The CAP Theorem, PACELC, and What Real Databases Actually Choose

The pick-two triangle is a misunderstanding: partitions happen regardless. PACELC, quorum arithmetic, and what DynamoDB, Spanner, and Postgres actually choose.

July 12, 2026Read
Connection Pooling: PgBouncer, RDS Proxy, and the Serverless Postgres Problem
ai-architecture1 min read

Connection Pooling: PgBouncer, RDS Proxy, and the Serverless Postgres Problem

A team raised max_connections from 100 to 5,000 and throughput fell. Connection pooling done right: PgBouncer modes, RDS Proxy, sizing math, and the serverless trap.

July 12, 2026Read
Multi-Cloud AI: Cost-Aware LLM Routing Without the Lock-In
ai-architecture1 min read

Multi-Cloud AI: Cost-Aware LLM Routing Without the Lock-In

An enterprise spent nine months duplicating its AI stack across two clouds; every real incident was a quota ceiling. Multi-provider routing: the architecture that pays.

July 11, 2026Read
LLM-as-a-Judge vs Deterministic Heuristics: Who Grades the Model?
ai-architecture1 min read

LLM-as-a-Judge vs Deterministic Heuristics: Who Grades the Model?

A judge scored 4.6/5 for months while users churned over wrong answers. When to use LLM-as-a-judge, when deterministic checks win, and the calibration that separates them.

July 11, 2026Read
Rust vs C++ for On-Device Inference Engines
ai-architecture1 min read

Rust vs C++ for On-Device Inference Engines

Memory safety versus bare-metal control for mobile AI engines. Where the three crash classes live, why ecosystem gravity beats safety, and where to put the unsafe seam.

July 10, 2026Read
The CACTUS Framework: Automated Data Quality Gates for RAG
ai-architecture1 min read

The CACTUS Framework: Automated Data Quality Gates for RAG

Your RAG hallucinates because the corpus is wrong, not the model. The six CACTUS data-quality gates, the heuristic-vs-judge cost split, and how to measure it.

July 10, 2026Read
Batch LLM Inference: Processing Millions of Documents Without Going Broke
ai-architecture1 min read

Batch LLM Inference: Processing Millions of Documents Without Going Broke

A classification backfill through the chat endpoint turned a $400 job into $3,000 and still failed. Batch LLM inference: half the cost, resumable, built for finite jobs.

July 10, 2026Read
PromptOps: Managing Prompts as Code Before They Break Production
ai-architecture1 min read

PromptOps: Managing Prompts as Code Before They Break Production

One prompt edit, shipped with no version history and no eval, silently regresses production for a week. PromptOps: prompts as versioned, tested, rollback-able code.

July 10, 2026Read
TOGAF vs Zachman: Do You Still Need an EA Framework in 2026?
ai-architecture1 min read

TOGAF vs Zachman: Do You Still Need an EA Framework in 2026?

TOGAF is a method, Zachman a taxonomy, and neither works as shipped. The decision table, where both fail, and the slim EA stack that survives contact with 2026.

July 9, 2026Read
Flutter and React Native at Scale: When Many Teams Share One App
ai-architecture1 min read

Flutter and React Native at Scale: When Many Teams Share One App

Three squads in one mobile codebase means merge queues, 45-minute CI, and blocked releases. Feature modules, monorepo tooling, and the tribe topology that fix it.

July 9, 2026Read
ACID Transactions and Isolation Levels, Explained with Failures
ai-architecture1 min read

ACID Transactions and Isolation Levels, Explained with Failures

Dirty reads, lost updates, and write skew: the bugs each isolation level permits, what Postgres and MySQL actually default to, and the cheapest weapon per write path.

July 8, 2026Read
Database Sharding vs Partitioning: Scaling Beyond One Box
ai-architecture1 min read

Database Sharding vs Partitioning: Scaling Beyond One Box

Partitioning makes one box manageable; sharding makes many boxes act as one database. The decision table, shard-key design, and the scaling ladder to exhaust first.

July 8, 2026Read
Loop Engineering for AI Agents: The Complete Guide
ai-architecture1 min read

Loop Engineering for AI Agents: The Complete Guide

"I don't prompt Claude anymore. My job is to write loops." The complete guide to loop engineering: brakes, context rot, tool contracts, and the verifier that says no.

July 7, 2026Read
Kafka vs RabbitMQ vs SQS: Choosing a Message Broker
ai-architecture1 min read

Kafka vs RabbitMQ vs SQS: Choosing a Message Broker

Kafka vs RabbitMQ dissolves once you name the workload: a log for facts many systems read and replay, a queue for jobs done once. The decision table and AI angle.

July 7, 2026Read
Database Indexing: How B-Trees Power Postgres and MySQL
ai-architecture1 min read

Database Indexing: How B-Trees Power Postgres and MySQL

The same query: 2ms or 3,000ms — the difference is one B-tree. How database indexes actually work, composite column order, covering indexes, and when the planner says no.

July 7, 2026Read
JWT vs Sessions: Authentication Architecture That Doesn't Bite Back
ai-architecture1 min read

JWT vs Sessions: Authentication Architecture That Doesn't Bite Back

The decade-old JWT vs sessions fight has a boring answer: both, deliberately. Short access JWTs, rotating refresh tokens, httpOnly cookies, and a real revocation story.

July 6, 2026Read
Background Jobs and Task Queue Architecture: BullMQ, Celery, and SQS
ai-architecture1 min read

Background Jobs and Task Queue Architecture: BullMQ, Celery, and SQS

The request cycle is for accepting work; the queue is for doing it. BullMQ vs Celery vs Sidekiq vs SQS, the five production requirements, and the AI-batch job shape.

July 6, 2026Read
04Contact

Bring this thinkingto your business.

Start a
conversation

One essay, most weeks. No noise.