How I Actually Use Claude Code as a Senior Engineer: Part 2
Fan out Agents, Real Scenarios for MCP, Mode Switching, Session Size
Aayush Ostwal
Gen AI Engineer | AI Agents | Building Scalable, event-driven, distributed applications for last 5 years
Live Last read · last published · next check
Latest posts
I Scaled Postgres with RDS Proxy in Production: Here’s What Happened
Hitting PostgreSQL connection limits? Learn how to use AWS RDS Proxy in production to multiplex connections, prevent database crashes during traffic spikes, and implement it using AWS CDK.
Thinking in Graphs: A Cypher Crash Course for SQL Engineers
5 min Crash Course
Datadog’s AI Report: 5 Things Senior Engineers Can’t Ignore
Why prompt engineering isn't enough anymore: The hidden costs, rising engineering debt, and new rules of production AI.
The 15 AI Agent Design Patterns Every Engineer Should Know (With Production Examples)
A field guide for engineers who want the right pattern, not the most impressive one.
How I Actually Use Claude Code as a Senior Engineer
Persistent memory, MCPs, project-scoped context, skills, and the workflows that genuinely save me time.
CLAUDE.md — the configuration file that makes Claude Code actually yours
CLAUDE.md — the one file that makes Claude Code actually yours
Your AGENTS.md File Is Making Your AI Coder only 3% better?
A rigorous new study from ETH Zurich and DeepMind shows that your carefully curated AGENTS.md file might be hurting your coding agent’s performance—and burning cash. Here’s what actually works.
Scaling Claude Code: How We Sync Custom AI Skills Across 100+ Repos
Github Submodules, Scripts, and NPM Package. See Implementation Now
Why Your AI “Agent” is Actually Just a Skill: Agent V/s Skill
Build BOTH. Together they create AI systems that are precise, purposeful, and powerful.
[Part 2] Context Engineering: Token & Memory Management in LLMs. Practical Guide for AI Engineers
What really happens inside an LLM after you hit “send”? From token explosions to context rot, uncover the unseen bottlenecks.
[Part 1] Context Engineering: 7 Prompt Engineering Methods
Reduced Hallucination, Get Accurate, Consistent, and Concise Results
Why “Just Coding” Will Get You Fired in 2026 (And What to Learn Instead)
What do Netflix, Meta, Amazon use? The tech bubble burst, and the rules changed. Here is the battle-tested, real-world roadmap to staying relevant in an age of AI and hyper-efficiency.