// THE STACK

Opinionated tools.
Intentional choices.

Every tool selected for a reason. Every integration tested in production. This is what "code-native finance" actually means.

⊞

Open Source First

Prefer tools that don't lock you in. DuckDB over Snowflake for most workloads. Evidence over Tableau. Your data, your models, your code.

⟁

Version Control Everything

If it can't live in Git, it probably shouldn't exist. Models, dashboards, pipelines—all tracked, diffed, and reversible.

◎

Boring Where It Counts

QuickBooks is fine. Gusto works. Don't overthink the foundation. Save the innovation for the analysis layer.

// Foundation Layer
The unglamorous stuff that makes everything else possible.

Clean books are a prerequisite, not a deliverable. These tools keep the foundation solid.

📚

QuickBooks Online

FOUNDATION

General ledger, AP/AR, bank reconciliation. The boring center of every financial stack.

Use case: Primary GL for clients under $20M revenue
👥

Gusto

FOUNDATION

Payroll, benefits, compliance. Handles state registrations, tax filings, and HR basics.

Use case: Payroll for US-based teams, 1-200 employees
💳

Ramp / Brex

FOUNDATION

Corporate cards with automatic categorization, receipt capture, and spending controls.

Use case: Expense management, department budgets
📄

Bill.com

FOUNDATION

AP automation, approval workflows, vendor payment scheduling. Beats manual check runs.

Use case: Vendor management, approval chains
// Analysis Layer
Where data becomes insight.

The modeling and transformation layer. Python and SQL working together.

🐍

Python + Pandas

ANALYSIS

The core modeling language. Operating models, cohort analysis, scenario engines—all in Python.

# Bottoms-up MRR proration
def prorate_mrr(contracts_df):
    return contracts_df.assign(
        prorated_mrr=lambda x: x.mrr * 
            x.days_active / x.days_in_month
    )
🦆

DuckDB

ANALYSIS

In-process SQL database. Handles gigabytes of data without a server. Perfect for financial modeling.

-- Fast aggregation, no infrastructure
SELECT cohort_month, 
       SUM(mrr) as total_mrr
FROM read_parquet('contracts/*.parquet')
GROUP BY 1
🔧

dbt

ANALYSIS

SQL transformations as code. Version-controlled, tested, documented. The backbone of the data pipeline.

Use case: Staging tables, marts, incremental models
🗄️

PostgreSQL

ANALYSIS

When you need a real database. Production analytics, shared data stores, API backends.

Use case: Persistent storage, multi-user access
// Intelligence Layer
AI and ML where it matters.

Machine learning for prediction, LLMs for augmentation. Not AI-for-AI's-sake.

🤖

Claude API

INTELLIGENCE

Natural language on your data. Automated variance commentary, anomaly explanations, data quality checks.

# AI-generated variance commentary
response = claude.messages.create(
    model="claude-sonnet-4-20250514",
    messages=[{
        "role": "user",
        "content": f"Explain: {delta}"
    }]
)
📈

scikit-learn

INTELLIGENCE

Classical ML for structured data. Churn prediction, revenue drivers, anomaly detection.

# Churn risk scoring
model = RandomForestClassifier()
model.fit(X_train, y_churn)
risk_scores = model.predict_proba(X_new)[:,1]
🚀

XGBoost

INTELLIGENCE

Gradient boosting for when accuracy matters most. Demand forecasting, driver analysis.

Use case: Revenue forecasting with many features
💻

Claude Code

INTELLIGENCE

AI pair programmer. Accelerates model development, automates boilerplate, catches bugs.

Use case: Development velocity, code review
// Visualization Layer
Dashboards built like software.

BI tools that deploy from Git, version control like code, and scale without manual work.

📊

Evidence.dev

VISUALIZATION

Markdown + SQL = production dashboards. Version controlled, deploys on commit, looks beautiful.

```sql monthly_arr
SELECT month, SUM(mrr) * 12 as arr
FROM contracts GROUP BY 1
```

<LineChart data={monthly_arr} x=month y=arr/>
⚡

Rill

VISUALIZATION

Blazing fast exploration. OLAP cube on your laptop. Perfect for ad-hoc analysis.

Use case: Exploratory analysis, fast iteration
🎯

Streamlit

VISUALIZATION

Python apps in minutes. Internal tools, scenario models, interactive reports.

Use case: Custom tools, what-if models
📈

Power BI

VISUALIZATION

When the client needs Microsoft. Enterprise integration, familiar interface.

Use case: Enterprise clients, Excel users
// Infrastructure
The invisible plumbing.

Deployment, version control, and orchestration.

🐙

GitHub

All code, all models, all dashboards

▲

Vercel

Dashboard deployments

☁️

Cloudflare

Static sites, edge caching

🔄

GitHub Actions

CI/CD, scheduled jobs

🐳

Docker

Reproducible environments

📦

uv + rye

Python package management

Ready to see this stack in action?

Let's talk about upgrading your financial infrastructure.

Schedule a Discovery Call