MJ
Manish Joshi
ServicesPortfolioFree AI ToolsBlogContact
Start Project →
MJ
Manish Joshi
ServicesPortfolioFree AI ToolsBlogContact
Start Your App →💬 Chat on WhatsApp (+91 95489 50280)
MJ
Manish Joshi

AI-Powered Mobile App Developer. Building production Flutter iOS & Android apps with integrated GenAI, LLMs, computer vision, and scalable ML backends.

Services

  • AI Mobile App Dev
  • Custom Flutter Apps
  • Add AI to Existing Apps
  • AI & ML Infrastructure

Work

  • Case Studies
  • Dliva Delivery
  • SnapQuote AI
  • About & Credentials

Resources

  • Free AI Developer Tools
  • Start Project
  • WhatsApp: +91 95489 50280
  • Privacy Policy

Built with by Manish Joshi

© 2026 manishjoshi.online · All rights reserved

Notion-Powered Engineering CMS

Engineering & Architecture Insights

Deep dives into Flutter mobile ecosystems, LLM pipeline design, LangGraph multi-agent systems, and production software architecture.

All Insights (27)Mobile App DevelopmentAI
AIAI & GENAI PIPELINES

Temporal.io for Fault‑Tolerant Distributed Workflows in Cloud‑Native Backends

Production InsightsManish Joshi
Sep 11, 2026• 5 min read

Temporal.io for Fault‑Tolerant Distributed Workflows in Cloud‑Native Backends

Temporal.io provides durable task queues and state‑machine semantics that let backend engineers build fault‑tolerant AI pipelines. The platform ensures workflows survive crashes, scale across regions, and simplify cloud‑native orchestration for distributed systems.

Manish JoshiRead Article
AIAI & GENAI PIPELINES

Beyond Chain-of-Thought: Implementing Stateful Multi-Agent Orchestration with LangGraph and Human-in-the-Loop Safeguards

Production InsightsManish Joshi
Sep 10, 2026• 6 min read

Beyond Chain-of-Thought: Implementing Stateful Multi-Agent Orchestration with LangGraph and Human-in-the-Loop Safeguards

AIAI & GENAI PIPELINES

The Linear Algebra Behind Transformer Self‑Attention: From QKV Projections to Efficient Implementations

Production InsightsManish Joshi
Sep 10, 2026• 6 min read

The Linear Algebra Behind Transformer Self‑Attention: From QKV Projections to Efficient Implementations

AIAI & GENAI PIPELINES

Fine-Tuning vs RAG: A Simple Way to Personalize AI With Your Own Data

Production InsightsManish Joshi
Sep 8, 2026• 6 min read

Fine-Tuning vs RAG: A Simple Way to Personalize AI With Your Own Data

Discover when to use Retrieval-Augmented Generation (RAG) VS Fine-Tuning to personalize AI models with your private data. Learn the cost, speed, accuracy, and architecture trade-offs for production LLM apps.

AIAI & GENAI PIPELINES

Do Google Searches Without AI Cause Global Warming? The Truth About Search, AI Search, and Energy

Production InsightsManish Joshi
Sep 6, 2026• 8 min read

Do Google Searches Without AI Cause Global Warming? The Truth About Search, AI Search, and Energy

← PreviousPage 2 of 2

LangGraph introduces a directed state graph that lets multiple LLM agents share context, pause for human approval, and resume with full audit trails. This stateful multi‑agent workflow replaces fragile sequential chaining with deterministic, inspectable orchestration.

Manish JoshiRead Article

This post breaks down the mathematics behind transformer self‑attention, showing how Q, K, and V projections turn into simple matrix multiplications. It also covers practical tricks for reducing the quadratic cost in real‑world deployments.

Manish JoshiRead Article
Manish JoshiRead Article

Do Google searches without AI cause global warming? Learn how traditional Google Search compares with AI search, what deep search AI actually does, how much energy AI queries use, and which AI search visibility metrics and KPIs matter for businesses in 2026.

Manish JoshiRead Article