Arnav Amal Ray
Platform-agnostic agentic AI for enterprise finance

I build platform-agnostic AI agents, accountable in enterprise finance.

An SAP S/4HANA finance consultant who is not locked into one vendor's AI. I build with SAP, Microsoft Copilot Studio, or a local LLM depending on what the problem and the data demand, and I prove every layer myself: retrieval, agents, integration, deployment, on-premise when data cannot leave the building.

Day Job

SAP S/4HANA finance consultant. I work where process, compliance, and technology meet, which is exactly where enterprises are now putting AI agents to work.

How I Think

I frame the problem, design the solution, and direct AI to build it. The decisions are mine. I own architecture, testing, deployment, and the result. AI handles execution speed.

Latest writing Why SAP API Policy v4 makes Clean Core a prerequisite for enterprise AI

Cross-platform

Across platforms, not inside one

I build the same patterns on whichever platform fits the problem: SAP, Microsoft Copilot Studio, or open-source and local LLMs. That means I understand what sits under every vendor's agents, not just one. When data cannot leave the building, I build it to run entirely on local hardware.

SAP S/4HANA and BTP Microsoft Copilot Studio Open-source and local LLMs

How I Work

From problem to production

The discipline is the same whether the problem is a broken finance process or a missing data pipeline, and whether I build it in SAP, Microsoft, or a local LLM. Define it clearly, then build the right thing, not just the obvious thing.

01
Frame the problem

Separate the stated problem from the real one.

02
Design the system

Define the architecture before touching a tool.

03
Build and validate

Direct AI to build it, then test against real conditions.

04
Ship, govern, and learn

Go live early, govern what runs, and treat first use as the start of learning.

Projects

Proof over theory

Each project is a complete end-to-end build: retrieval, an agent that acts, an integration pipeline, and an oversight layer. Individually each solves a real problem. Together they are one system, the building blocks of agentic enterprise finance.

Learning in PublicLocal · On-Premise

Local LLM (RAG) for Immigration Questions

Retrieval and grounding: on-premise AI that answers only from cited source text and never invents it.

Local LLMRAGPython

The Problem

Standard AI assistants either refuse to answer legal questions or invent rules that do not exist. German immigration law spans multiple statutes with hundreds of cross-referenced sections; a single eligibility question can require reading five different parts of the law at once.

What I Built

A system that runs entirely on local hardware, so no data ever leaves the machine, GDPR compliance by design. It retrieves the most relevant legal sections for any question and generates an answer grounded only in those sections, citing exactly where each point comes from.

Enterprise parallel: grounded assistants over SAP, finance, and compliance documentation, on-premise where data cannot leave the building.

Actively UsedLive · Serverless

Family Finance and Goals AI Bot

An agent that takes action: a conversational bot that turns a message or a receipt photo into structured data.

TelegramGoogle SheetsServerless

The Problem

Logging a family expense means opening an app, finding the right category, and typing it manually. The friction is small enough to skip, until weeks of spending disappear. Goals suffer the same way: written down once, never revisited.

What I Built

A Telegram bot that reads a message or a photo of a receipt and writes the expense directly into a shared Google Sheet in under a second. Goals work the same way: one natural-language message creates a tracked entry. Zero apps, zero manual effort.

Enterprise parallel: conversational agents that capture transactions and post them into enterprise systems.

Learning in PublicOpen Source · Python

Portfolio Rebalancer

An integration pipeline: it turns a broker PDF into priced holdings, risk metrics, and suggested actions.

PythonYahoo Finance APILLM

The Problem

An ETF portfolio drifts away from its target allocation as prices move, but the holdings sit locked inside a broker PDF, and the maths to quantify the drift, and the risk that comes with it, is tedious and error-prone to do by hand.

What I Built

A Python tool that parses a broker statement into holdings, resolves each position to a live price from Yahoo Finance, computes risk metrics such as Sharpe, Sortino, max drawdown and VaR, then reports the buy and sell trades that would close the gap to target, with an optional LLM narrative. It is educational only: it never places trades and is not investment advice.

Enterprise parallel: document-to-decision pipelines that read statements and surface actions for finance teams.

Concept DemoLive · Multi-System

TreasuryFlow - Governance and Visibility

The oversight layer: live AP, AR, and cash positions across systems, in one view.

API integrationDashboard

The Problem

Once agents and pipelines are acting on finance data, the missing piece is oversight, one trustworthy view of what is happening across systems, so a human can see and own the result.

What I Built

A personal concept demo that pulls AP, AR, and cash positions via API into one live dashboard. It is the oversight layer over the other three building blocks. It is a personal concept demo, separate from any client or employer work.

Enterprise parallel: the governance and visibility layer over a system of finance agents.

Where this is going

The tools to build agents are arriving fast, from SAP and every other vendor. As they get more autonomous, the scarce role is not building one. It is understanding the finance consequence well enough to set the boundaries, prove the output is correct before it reaches the books, and own the result. That is where I focus.

Let’s connect

Whether you are building an AI-enabled product, thinking about how AI changes enterprise finance, or looking for someone who bridges both worlds.