Skip to content
View mlaharsh's full-sized avatar
  • ETHARA AI
  • GURUGRAM

Block or report mlaharsh

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mlaharsh/README.md

Harsh Verma

Financial Analyst · Data-Driven Decision Support

Turning customer, product, and operational data into margin, retention, and forecasting decisions.

Snapshot

  • Role: Financial Analyst at Ethara AI
  • Location: Gurugram, India
  • Focus: Unit economics · Margin optimization · Churn & retention · Demand forecasting · Variance analysis
  • Toolkit: Python (pandas, NumPy, scikit-learn) · SQL · Advanced Excel · Tableau · Streamlit · Git

What I build

I like the point where analysis stops being a spreadsheet and starts being a decision. My projects pair a reproducible model with an interface someone can actually use — a scorer, a dashboard, or a workbench — and document the assumptions so the numbers can be challenged.

Project Question it answers Stack
Amex Premium Card Tier Migration & Margin Optimization · live scorer Which card tier maximizes annual margin for each customer — upsell, maintain, downsell, or retain? Python · scikit-learn · Streamlit · GitHub Pages
Customer Churn Prediction · live showcase Which customers are likely to leave, and where should retention effort go? Python · scikit-learn · Streamlit · Docker
Supply Chain Inventory Optimization · live dashboard How much inventory is enough — balancing service level, stockout risk, and working capital? Python · SQL · forecasting · Tableau

How I approach analysis

  1. Define the decision first. A model is only useful if it changes what someone does on Monday.
  2. Keep it reproducible. Versioned pipelines, pinned dependencies, tests, and a written model card.
  3. Show the uncertainty. Scenario ranges and sensitivity — not just a point estimate.
  4. Make it legible. Every output has a plain-English interpretation and a stated limitation.

Toolbox

Python SQL pandas scikit-learn Tableau Streamlit Excel Git

Currently

  • Building decision tools around card economics, retention, and inventory planning.
  • Sharpening SQL and financial modeling depth; reading more on forecasting and causal inference.
  • Open to conversations on analytics, FP&A, and applied machine learning in finance.

Connect

GitHub

Every project here is documented, tested, and reproduced from a clean clone. If a number matters, it should be traceable.

Popular repositories Loading

  1. Customer_Churn_Prediction Customer_Churn_Prediction Public

    Customer churn prediction and retention workbench: a reproducible scikit-learn pipeline, Streamlit app, model card, tests, and public showcase.

    Jupyter Notebook

  2. supply-chain-inventory-optimization supply-chain-inventory-optimization Public

    End-to-end inventory optimization: ABC-XYZ, EOQ, safety stock/ROP, demand forecasting, Monte Carlo stockout risk, and Tableau dashboard.

    Python

  3. amex-tier-migration amex-tier-migration Public

    Premium card tier migration & margin optimization: churn model, tier economics, Streamlit workbench, and a live client-side scorer.

    Python

  4. mlaharsh mlaharsh Public

    GitHub profile - financial analytics portfolio