Popular works
ThreadPoolExecutor vs ProcessPoolExecutor (Practical Guide)
A practical, student-friendly explanation of concurrency and parallelism in Python, with real-world examples using ThreadPool...
Why the Command pattern works so well for backtesting
Do you know design patterns? One major category is Behavioral patterns. Command is part of it, and I will explain it with a r...
Caching Fundamentals for dummies like me
A clear and student-oriented explanation of caching fundamentals, cache states, eviction strategies, and architectural trade-...
All articles
Basic State Machine Pattern for dummies like me
When I was a student, I didn’t understand state machines. Not because they are hard, but because they are usually explained badly. This article fixes that.
The Registry Pattern - How not to lose your data in the Async void
Managing one session is easy. Managing hundreds of concurrent WebSockets, each with its own state, analytics, and market data without causing a memory leak or a race condition? That's where the Registry Pattern saves your life.
Structured SQL Path for RAG: Planner, Command Bus, and Safe Execution
Implement a safe structured SQL path with planner contracts, command bus handlers, and guardrails for deterministic production execution.
Smart Routing for RAG: Choosing Between BM25, SQL, and HYBRID
Design a smart router that chooses BM25, SQL, or hybrid retrieval using confidence thresholds, margins, and latency-aware decision rules.
Query Normalization and Time Intelligence for Trading RAG
Normalize noisy user queries and resolve time expressions reliably so routing and SQL generation stay consistent in real trading workflows.
Retrieval Foundation for Custom RAG: Postgres FTS + Local TF-IDF
Build a practical retrieval backbone with Postgres full-text search and local TF-IDF, including chunking strategy and ranking behavior.
Domain Modeling for RAG in Trading Journal Platform: Trades, Strategies, Tasks, Users
How to design a retrieval-safe domain model for finance RAG with clear entities, ownership boundaries, and queryable business semantics.
Why We Built a Custom RAG for Finance Trading Journal Platform
Why LLM-only assistants fail for Trading Journal Platform, and how a custom RAG architecture improves trust, accuracy, and production safety.
The Bloom Filter – Fast membership checks with bounded error
A Bloom filter is a compact data structure that answers membership questions quickly: either an element is definitely not in the set, or it may be in the set. This article explains the idea from first principles, then shows a small C implementation and practical Typescript use cases.
Understanding JSONB in PostgreSQL through the lens of Performance and Flexibility
I propose a deep dive into the JSONB format in PostgreSQL. Is it just a storage trick or a real architectural shift? Let's explore why this binary format is a must-have for modern backends.