Deep Dives
System design, AI engineering, and business finance — complex topics explained simply with interactive visuals.
Business Finance
Reading strategy from financial statements — real companies, real numbers
Netflix's Three Reinventions
(and What They Cost)
From $36M shipping DVDs to $31.6B making Squid Game. Each reinvention left a fingerprint on the financial statements — if you knew where to look.
Bahgat Ahmed
·
November 2025
·
30 min read
How to Read a Company's Strategy
From Its Numbers
The 4-step framework that Wall Street analysts use to decode any company's strategy from its financial statements.
Bahgat Ahmed
·
December 2025
·
25 min read
Software
System design, scaling patterns, and production engineering
The Complete Software Map
What Actually Happens When You Click a Button
Follow a single user click through every layer of a real system. From network to database, from security to deployment. The complete picture.
Bahgat Ahmed
·
August 2025
·
55 min read
From Weekend Project
to Production
What actually matters at each stage—and what doesn't. A complete guide to PoC → MVP → Production with checklists, decision frameworks, and AI-specific traps.
Bahgat Ahmed
·
September 2025
·
25 min read
The 95% Problem:
Understanding DB Connections
Your system works 95% of the time, but randomly fails with "connection closed". Learn why database connections fail, how connection pooling works, and why idle timeouts are the silent killer.
Bahgat Ahmed
·
September 2025
·
20 min read
AI
LLM engineering, memory patterns, and building AI applications
What Changed?
How Modern LLMs Evolved Beyond the Original Transformer
RoPE, RMSNorm, GQA, SwiGLU, Flash Attention, MoE — every upgrade from the 2017 transformer to today's models, explained by what it replaced and why.
Bahgat Ahmed
·
March 2026
·
30 min read
RAG & Agents:
How LLMs Find Answers & Take Action
From keyword search to autonomous agents — retrieval pipelines, vector databases, chunking strategies, tools, planning, memory, and the safety guardrails that keep AI from going rogue.
Bahgat Ahmed
·
October 2025
·
20 min read
Is It Actually Working?
Why "94% accuracy on the benchmark" almost hospitalized patients, why self-correction makes models worse, and how to build evaluation that catches what benchmarks miss.
Bahgat Ahmed
·
November 2025
·
22 min read
When RAG Isn't Enough
Your RAG retrieves the right documents, but the model still gets the answer wrong. Sometimes the model needs to actually learn your domain — and that changes everything.
Bahgat Ahmed
·
December 2025
·
20 min read
AI Memory:
From Stateless to Context-Engineered
Your chatbot forgets the user's name on message 48. A triage bot drops "allergic to penicillin" on message 12. These aren't edge cases — they're what happens when memory is an afterthought.
Bahgat Ahmed
·
January 2026
·
25 min read
Teaching Agents to Remember
Your AI agent forgets preferences, overwrites history, and can't answer "what changed?" From knowledge graphs to temporal awareness — the complete guide to agent memory.
Bahgat Ahmed
·
February 2026
·
35 min read
Get New Posts
Get notified when I publish new articles.
What brings you here?