PortfolioFinance · Analytics · AI

Open to full-time roles · Graduating Dec 2026

Deepak Raghunath

I turn financial data into decisions. 4.5+ years of P&L reporting, month-end close, controls and forecasting at Ameriprise Financial and Accenture. Now I build forecasting models and AI-assisted workflows with SQL, Python, Power BI and LLMs.

See selected work Run the live models
Based
St. Louis, MO
Studying
M.S. Business Analytics (FinTech Analytics, STEM), WashU Olin, Dec 2026
Previously
Ameriprise Financial · Accenture
Toolkit
SQL · Python · Power BI · Excel/VBA · Tableau · LLMs
RetireSim · live Open in the lab →
 
Paths that last 30 yrs
 
Median ending value
 
Without guardrails
5,000 simulated paths · $1M · 5% initial withdrawal · guardrails on Illustrative model, not advice

01Impact

Measured results, with context.

Each figure comes from my work in finance operations at Ameriprise Financial and Accenture, shown next to how it was achieved and where.

1,400+ hrs/yr

Reporting time saved every year

Self-serve Power BI dashboards and KPI frameworks, adopted org-wide for real-time leadership visibility.

Ameriprise · 2023–25
40% faster

Month-end close turnaround

SQL queries written with IT feed ledger data into Excel VBA macros across 25 of 45 close processes, with 95% SLA adherence maintained.

Accenture · 2021–22
$250K+

Financial discrepancies resolved

P&L and reconciliation reporting in Power BI, used across audit and remediation. Exception-checking rules prevented a further $200K+ in potential errors.

Accenture · 2021–22
100%

RMD compliance, no IRS penalties

Exception-monitoring dashboard with automated age-based checks. Try a synthetic version →

Ameriprise · 2023–25
+20% YoY

On-time completion of portfolio reviews

An auto-refreshing control dashboard flagged at-risk deadlines across 25+ advisor portfolio reviews a week, protecting advisory-fee revenue.

Ameriprise · 2023–25
30+ SOPs

Close and reporting procedures written

Paired with cross-training that saved $8,000 a year through less error rework and faster onboarding.

Accenture · 2021–22

2025 · Ameriprise Financial

Ammy Award

Operational Excellence, selected from 200+ members of the organization.

2021 · Accenture

Best People Award

Given to the top 5% of Accenture Operations employees.

02Selected work

Four problems, and how I worked them.

For each case: what was at stake, what I did myself, and where it stands now. Employer and client data stay private.

01ShippedAmeriprise Financial · Senior Financial Advisory Analyst · 2023–2025

Real-time visibility for advisory leadership, with deadlines flagged before they slip.

The context

Advisors were running 25+ portfolio reviews a week, and leadership needed a real-time view of them. Late reviews put advisory-fee revenue at risk, and a missed required minimum distribution (RMD) brings IRS penalties.

What I did

  • Built self-serve Power BI dashboards and KPI frameworks, later adopted org-wide.
  • Built an auto-refreshing control dashboard that flags at-risk review deadlines.
  • Built an RMD exception monitor that runs automated age-based checks.
Power BIDAXKPI designException rulesExcel
02ShippedAccenture · Premium Accounting Analyst · 2021–2022

Cutting month-end close turnaround by 40% while keeping controls in place.

The context

The team ran 45 month-end close processes against service-level targets. Each close meant ledger pulls, variance analysis, reconciliations and close packs, and every error meant rework.

What I did

  • Owned 25 of the 45 processes, covering variance analysis, reconciliations and close packs.
  • Wrote SQL with IT to feed ledger data into Excel VBA macros.
  • Built P&L and reconciliation reporting in Power BI, with exception-checking rules.
  • Wrote 30+ SOPs and led cross-training. Served as IT point of contact for a 50+ person team.
SQLExcel VBAPower BIReconciliationsVariance analysis
03DeliveredB&N Inspection · Data Analyst Intern, Financial Data & Reporting · Summer 2026

Eleven years of inspection PDFs turned into one source of truth, with an LLM doing the extraction.

The context

Eleven years of inspection data was fragmented across client sites, and much of it sat in unstructured inspection PDFs. Without a single source of truth, KPI reporting and forecasting were out of reach.

What I did

  • Consolidated the fragmented data into a single source of truth across client sites.
  • Built Claude (LLM) workflows that classify and geolocate assets from the inspection PDFs.
  • Built a VBA-driven Excel intake workbook with cascading validation that feeds the projection models.
Claude LLMDocument extractionGeolocationExcel VBAData modeling
04In progressBunge · WashU Olin CEL Practicum · Student Consultant · Aug 2026–present

Can retail scanner data predict demand for vegetable fats and oils?

The question

Bunge is a Fortune 500 agribusiness and food-ingredients company. We are testing whether Nielsen retail sales data can serve as a leading indicator of food manufacturers' demand for Bunge's vegetable fats and oils, to sharpen demand forecasting and planning.

My workstream

  • Cleaning and merging Nielsen retail data with Bunge demand data in SQL and Python, aligning products, categories and time periods.
  • Building a forecasting model that converts retail sales signals into projected ingredient demand.
  • Using Claude (LLM) to speed up the analysis.
SQLPythonpandasForecastingNielsen

03Decision lab

Don't take my word for it. Run the models.

Four models I built, running entirely in your browser. You can change every input, the assumptions are written next to each model, and the same inputs always give the same answer.

RetireSim

How long does a retirement portfolio last, and what do spending guardrails buy you? The model runs 5,000 simulated market paths and re-runs every time you change an input.

Guyton-Klinger guardrails

Annual steps, with each withdrawal taken at the start of the year. Returns are lognormal with the mean and volatility you set. The guardrails work like this: skip the inflation raise after a losing year if the withdrawal rate is above where it started; cut spending 10% if the rate is 20% above its start (except in the final 15 years); raise it 10% if the rate is 20% below its start. The random seed is fixed. Illustrative only, not financial advice.

Success rate
 
Median ending value
 
nominal dollars
10th pct ending
 
bad-luck outcome
Lowest real income
 
median path, vs year 1
5th–95th percentile25th–75thMedianSample paths

View percentile table

04Now

What I'm working on this fall.

Alongside the Bunge practicum, my time goes to a startup, a research panel and teaching. I've listed my own contribution to each, not the team's.

Jul 2026 – present · Startup

Ainstein

Co-Founder & CFO

Ainstein is a game-like platform for people who barely use AI yet. I set the product direction and commercial strategy, and brought together the four-person product, design and engineering team.

I built a 36-month driver-based financial model with Bear, Base and Bull scenarios. It covers the revenue build, unit economics, P&L, burn, runway and use of funds, and guides pricing and fundraising decisions.

Driver-based modelUnit economicsScenarios

Aug 2026 – present · Research

RBI banking regulation study

Research Assistant, WashU

I'm building a 1972–1998 district-level panel from digitized Reserve Bank of India Basic Statistical Returns. The study measures how RBI regulation changed where bank deposits, branches and credit went.

My work includes district and state boundary crosswalks that keep 25+ years comparable, cleaning and validating the panel in Stata, and dating regulatory changes from RBI circulars.

StataPanel dataCrosswalksEconometrics

Aug 2026 – present · Teaching

Graduate Teaching Assistant

WashU Olin

I hold virtual office hours and grade SQL and Tableau labs and homework, calibrating marks with the other TAs. The courses are Database Design & SQL, Data Visualization, and Big Data & Cloud Computing.

Before this, as a CFAR research assistant, I standardized finance-faculty databases across roughly 150 business schools.

SQLTableauSpark / Databricks

05Experience

The full record.

All my roles, newest first. The résumé has the complete detail. Download PDF →

Graduate, startup & research

Aug 2026 – Present

Student Consultant, Demand Forecasting · Bunge (WashU Olin CEL Practicum)

Testing whether Nielsen retail sales data leads food manufacturers' demand for Bunge's vegetable fats and oils. I merge the data in SQL and Python and build the forecasting model.

St. Louis, MO
Jul 2026 – Present

Co-Founder & CFO · Ainstein

Product direction, commercial strategy and team assembly. I built the 36-month driver-based financial model with Bear, Base and Bull scenarios.

St. Louis, MO
Mar 2026 – Present

Research Assistant & Graduate Teaching Assistant · Washington University in St. Louis

  • Banking Regulation Study (Aug 2026 – present): RBI district-level panel covering 1972–1998, with boundary crosswalks and Stata validation.
  • Graduate TA (Aug 2026 – present): Database Design & SQL, Data Visualization, Big Data & Cloud Computing.
  • Wells Fargo Advisors CFAR (Mar – Jul 2026): standardized finance-faculty databases across roughly 150 business schools.
St. Louis, MO
Jun 2026 – Aug 2026

Data Analyst Intern, Financial Data & Reporting · B&N Inspection

Consolidated 11 years of inspection data into a single source of truth. Built Claude LLM workflows to classify and geolocate assets, and a VBA intake workbook that feeds the projection models.

St. Louis, MO

Finance & analytics

Jan 2023 – Aug 2025

Senior Financial Advisory Analyst · Ameriprise Financial

Power BI dashboards and KPI frameworks that saved 1,400+ hours a year. A control dashboard that lifted on-time review completion 20% YoY. RMD exception monitoring that achieved 100% compliance. Ammy Award 2025.

Gurugram, India
Mar 2021 – Dec 2022

Premium Accounting Analyst · Accenture

Owned month-end close for 25 of 45 processes and cut turnaround 40% with SQL→VBA automation. Resolved $250K+ in discrepancies, wrote 30+ SOPs, and served as IT point of contact for a 50+ person team. Best People Award 2021.

Noida, India

06Models & side builds

Things I built to learn by doing.

Financial models and AI tools I built on my own time. Each one is labeled with what it is.

Excel model

Three-statement model: data-center business

An integrated income statement, balance sheet and cash-flow model for a capital-intensive data-center business. Revenue is built from contracted compute capacity (MW/GW), which drives the capex and depreciation schedule and a debt and financing waterfall. Scenario toggles included.

Web app · in the lab above

RetireSim Monte Carlo planner

A 5,000-path Monte Carlo projection engine with Guyton-Klinger guardrails and scenario fan charts. I'm extending it into a full-stack React/TypeScript planner with LLM-generated guidance.

AI agent · self-hosted

Morning Brief agent

An n8n agent that pulls Financial Times RSS feeds, uses Gemini 2.5 Flash to rank stories for fintech relevance, and emails a categorized morning brief through Gmail OAuth. Runs in Docker.

Prototype

DataPulse KPI builder

Turns a raw CSV into a KPI dashboard. It detects fields automatically, generates insights, checks for outliers and summarizes correlations, so people without a technical background can use it.

07About

Finance first. Builder by habit.

I'm a finance person who learned to build the tools that finance teams need.

I grew up in Delhi and studied financial markets at Jai Hind College in Mumbai. I then spent four and a half years in finance operations: first owning month-end close at Accenture, then building the dashboards and controls that advisory teams at Ameriprise relied on every week.

At WashU Olin I'm adding forecasting, econometrics and AI to that foundation. I work on the Bunge engagement, a research panel on Indian banking regulation, an AI-learning startup, and TA roles in SQL and data visualization.

Outside work: Liverpool FC, badminton, the gym (since 2017), and long road trips with a well-made playlist.

Education

2025 – Dec 2026

M.S. Business Analytics, Financial Technology Analytics

Washington University in St. Louis, Olin Business School · STEM-designated. Coursework: Advanced Corporate Finance, Database Design & SQL, Investment Theory, Options & Futures, Fixed Income.

2017 – 2020

B.Com, Financial Markets · GPA 4.0 / 4.0

University of Mumbai, Jai Hind College.

Toolkit

FP&A & finance
Budgeting, forecasting, driver-based and three-statement modeling, variance analysis, P&L bridges, scenarios, unit economics, month-end close, reconciliations.
Data & BI
SQL (Spark SQL / Databricks), Python (pandas, NLP), Power BI (DAX), Tableau, Excel (VBA, Power Query), Stata.
AI & automation
LLM workflows with Claude and Gemini, document extraction, n8n agents, prompt design.

08Contact

Let's talk about the numbers behind your next decision.

Emaildeepakraghunath1999@gmail.com LinkedIn ↗in/deepak-raghunath Phone+1 314-309-9932 Résumé ↓Download PDF
© 2026 Deepak Raghunath · St. Louis, MO Updated September 2026 · Lab models are illustrative, not investment or tax advice.