Asking the same question to SWE-1 by Windsurf, Claude 3.7 Sonnet thinking, openAI and Gemini 2.5, I'd like o assess and compare these three LLMs. The question is generic and simple "if i want to build up a codebase for garage parking system, please sketch our blueprint md file first, then write out codes" SWE … Continue reading LLM Model Comparison Using Garage Parking System Designing

We often hear people say, “I’ll try if things work out,” or “I’m not sure I’m good enough for this.” These statements aren’t reflections of facts; they’re reflections of fear. At the core of that fear lies self-doubt—a mental parasite that erodes potential long before failure ever happens. Let’s be clear: You absolutely cannot allow … Continue reading The Only Thing You Control – Your Own Mindset

Scheduling tasks to run at specific intervals or during certain conditions is a common requirement for many applications. wo popular approaches are using Windows built-in Task Scheduler versus leveraging APScheduler, a Python-based scheduling library. Let’s compare these two approaches by taking a look at two examples: Using APScheduler to define more sophisticated scheduling logic with … Continue reading Comparison of Two Approaches for Scheduling Tasks: Windows Scheduler vs. APScheduler

Creating UI is a necessary element in DAG workflow. here is the additional strucuture: index_management_ui/├── app.py # The main Flask application├── ui_database.json # A simple file to act as our database└── templates/├── layout.html # Base HTML template with Bootstrap CSS├── dashboard.html # The main monitoring dashboard└── details.html # The approval/rejection page for a single run … Continue reading Breakdown 4: UI Element of the DAG Workflow

airflow_project/├── dags/│ └── index_rebalancing_dag.py # Our main DAG file├── include/│ └── calculations/│ │ └── engine.py # Python code for the heavy lifting│ └── utils/│ └── reporting.py # Code for generating reports└── config/└── indices/└── sp500_rules.yaml # Our version-controlled index definition The workflow would look like graph TDA[get_index_config] --> B[collect_and_validate_market_data];B --> C[calculate_preliminary_rebalance];C --> D[generate_analyst_report];D --> E[notify_committee_for_review];E --> … Continue reading Breakdown 3: DAG Airflow Concrete

For complex scheduling tasks use Airflow, features are Scheduling: Uses cron-style scheduling Error Handling: Built-in retries and failure callbacks Dependencies: Clear task dependencies Modularity: Separates different concerns into tasks Configuration: Uses Airflow Variables and Connections Monitoring: Built-in UI for monitoring and logging from airflow.utils.dates import days_ago from airflow.operators.email_operator import EmailOperator with DAG( 'monthly_rebalancing', default_args=default_args, description='Monthly … Continue reading Breakdown 2: DAG AirFlow Generic

When the inventory grows, managing various versions with version control is essential. Applying "Version-controlled configuration files (YAML/JSON) for all index rules" is a solution. The fundamental shift is to stop treating index rules as descriptions in a Word document or entries in a database GUI, and instead treat them as source code. This practice is called Configuration … Continue reading Breakdown 1: Apply “Version-controlled configuration files (YAML/JSON) for all index rules”

A blueprint describes the architectural principles and best practices that a modern, large-scale index provider either uses or is actively migrating towards. The Core System Components (Immutable audit logs of all calculations) Workflow Orchestration Uses tools like Apache Airflow, Prefect, or custom schedulers Defines DAGs (Directed Acyclic Graphs) for each rebalance event Implements dependency management … Continue reading Blueprint of a Workflow on the Platform

Based on what I've seen, I totally vibe with this engineer's take on how he uses AI: The keys to do good prompting are: be very specific! provide visual context, screenshot don't do too much at once, do small iterations

Big asset managers like BlackRock, Vanguard use custom-built schedulers, cron-based orchestration, enterprise ETL tools, and monolithic job control systems to manage workflow. These setups were often brittle and difficult to manage at scale but were the standard in the 2000s and early 2010s. Finance had complex, interdependent systems decades before open-source orchestration tools existed. Regulations … Continue reading Apache Airflow and Prefect