Product Manager · San Francisco

"From zero to one—and beyond."

Hello World.

I'm Tejas Prakash Patil.

I'm a human being and a product manager based out of San Francisco.

I've spent more than a decade building thoughtful, reliable, and scalable software across a variety of companies.

I genuinely love building products — not just shipping features, but solving problems that actually matter to people. I've learned that the best products come from really understanding the problem, not just executing on a spec.

I'm a self-starter who's comfortable with ambiguity, works well with engineers and designers, and isn't afraid to push back when something doesn't feel right. Lately, I've been going deep on Generative AI and LLMs — genuinely excited about where it's all heading and how it changes what's possible in product.

I'm actively looking for new opportunities to have a meaningful impact on the world. If you'd like to collaborate, email me at tejas.p29@gmail.com.

Work

Projects

Industry Experience

IP:01

Sam's Club (Walmart)

App Merge — Scan & Go + eCommerce

Led the integration of Sam's Club's Scan & Go and eCommerce mobile apps into a single unified application, delivering 100+ features to millions of users and improving customer engagement by 30%.

iOSAndroidFirebaseAgile

IP:02

Sam's Club (Walmart)

Same-Day Delivery

Launched Sam's Club's Same-Day Delivery channel end-to-end — from discovery to checkout and fulfillment — generating over $100M in revenue and integrating the Sam's Cash rewards system.

Product StrategyAPIsA/B TestingAWS

IP:03

Sam's Club (Walmart)

Retail Media Ads Platform

Built Sam's Club's mobile advertising platform from zero, becoming one of the largest profit streams in the eCommerce business and delivering $100M+ in profit.

Ad TechData ModelsREST APIsFirebase

IP:04

Sam's Club (Walmart)

AI-Powered Product Discovery

Drove the shift from keyword search to conversational, agentic discovery — partnering with Google's Gemini AI to surface products naturally when users express intent, turning shopping into a conversation.

Generative AIGeminiLLMsPersonalization

IP:05

Glooko

4.0 App Relaunch

Contributed to the modernization of Glooko's diabetes management platform — a ground-up relaunch that improved how patients and physicians track, share, and act on diabetes data.

iOSAndroidHealthcareData Pipelines

IP:06

Royal Bank of Canada

Mobile Banking App Update

Helped build RBC's next-generation mobile banking app across Android, iPhone, and BlackBerry — raising the quality bar for millions of banking customers through rigorous automation and testing.

AndroidiOSBlackBerryFinTech

Lab

Personal Projects

Side Projects & Experiments

PP:01

Personal / ML

Weather Temperature Predictor — Multiple Linear Regression

A supervised machine learning model built in Python to predict temperature (°C) from real atmospheric inputs. Trained a Multiple Linear Regression model using scikit-learn, serialised with pickle, and deployed as a live Flask REST API. Forecasting temperature from weather data has real-world impact across agriculture, energy planning, and daily logistics.

Problem Type

Supervised Learning
Regression

Algorithm

Multiple Linear
Regression (MLR)

Target Variable

Temperature (°C)

Features (4)

Humidity (%) · Wind Speed (km/h)
Pressure (hPa) · Rainfall (mm)

Python scikit-learn Flask Multiple Linear Regression Supervised ML REST API pickle

View on GitHub →

PP:02

Personal / ML

Polynomial Regression Explorer

An interactive supervised learning demo that fits a polynomial regression curve to a synthetically generated dataset. Enter any degree from 1 to 10 and watch the curve update live — a great visual for understanding underfitting (degree 1) versus overfitting (degree 9+). Built with Python, scikit-learn, and Flask, rendered in the browser via Chart.js.

Problem Type

Supervised Learning
Regression

Algorithm

Polynomial Regression
(user-defined degree)

Dataset

Synthetic — 200 points
0.8x² + 0.9x + 2 + noise

Key Concept

Bias–Variance tradeoff
Underfit vs Overfit

Python scikit-learn Flask Polynomial Regression Supervised ML Chart.js REST API

View on GitHub →

PP:03

Personal / ML

Titanic Survival Classifier — Binary Classification

An interactive supervised learning demo that predicts whether a Titanic passenger would have survived, based on their profile. Enter passenger details — gender, age, class, fare, port, and family size — and a logistic regression classifier trained on the 891-passenger RMS Titanic dataset returns a survival prediction with probability score and factor-by-factor breakdown. A hands-on illustration of binary classification, feature engineering, and the sigmoid function.

Problem Type

Supervised Learning
Binary Classification

Algorithm

Logistic Regression
σ(z) = 1 / (1 + e⁻ᶻ)

Dataset

RMS Titanic — 891 passengers
38.4% survival rate

Features (7)

Sex · Age · Pclass · Fare
SibSp · Parch · Embarked

Python scikit-learn Logistic Regression Binary Classification Supervised ML Feature Engineering JavaScript

Live Prediction Demo

Would you have survived the Titanic?

Gender

♀ Female ♂ Male

Age

Ticket Class

▾

Ticket Fare (£)

Embarked From

▾

Siblings / Spouses

Parents / Children

View on GitHub →

I'm Tejas Prakash Patil.

Projects

App Merge — Scan & Go + eCommerce

Same-Day Delivery

Retail Media Ads Platform

AI-Powered Product Discovery

4.0 App Relaunch

Mobile Banking App Update

Personal Projects

Weather Temperature Predictor — Multiple Linear Regression

Polynomial Regression Explorer

Titanic Survival Classifier — Binary Classification

Resume

Education

Certifications

Socials