Bharath SrivatsanSoftware
Work
I am currently a senior staff engineer and the product architect for Stripe's front door integrations (Checkout, Billing, and Payments). Previously, I was an engineering lead and founding engineer for Stripe Agentic Commerce, Stripe Orchestration, Stripe Treasury, and other products.
I specialize in API design, having shaped the latest evolution of our checkout APIs, our Payment Methods APIs, our agentic commerce APIs, and our foundational money movement APIs. I've also reviewed over 50 proposals over six years as an API reviewer, for products as wide-ranging as Radar, Terminal, and Billing.
I am an eight-time keynote speaker at marquee Stripe events, including Stripe Sessions and Convergence. I designed product demos for all eight of these conferences and a host of other company all hands and tech fora.
If you'd like to chat APIs, zero to one product development, or demo design, shoot me an email!
Patents
Generating Tokens for AI Agents (Token Workflow)
End-to-end constrained access token generation for AI agent payment sessions, with merchant-scoped constraints and direct execution. Filed 2025.
Generating Tokens for AI Agents (Virtual Card Conversion)
Token-to-virtual-card conversion path for remote execution, generating temporary credentials with inherited constraints. Filed 2025.
Generating Tokens for AI Agents (Fraud Scoring)
Two-stage fraud scoring for AI agent sessions with score drift detection between token generation and execution. Filed 2025.
Detection and Blocking of Unauthorized AI Agent Operations
Privacy-preserving detection of unauthorized operations in AI agent payment channels with configurable step-up authentication. Filed 2025.
Predicting Network Operations Initiated by Automated Processes
Detecting bot and automated abuse of AI agent payment channels, distinguishing non-human activity from legitimate agent use. Filed 2025.
Asynchronous Scheduling and Processing of Network Events
ML-driven async payment processing with two-stage pipeline for success prediction and optimal retry timing. Filed 2025.
Papers
Understanding Users Through Targeted Ads
Senior Thesis. Conducted the largest live-user study of targeted ads (at the time). Co-winner of the Princeton CS dept's 'Outstanding Thesis Award'.
Personally Stylized Post-Processing
Models that learn the stylistic preferences of individual photographers and post-process new images based on these unique styles. Python and Tensorflow.
Synonym-Based Memory Networks
A new flavor of Memory Network to improve performance on modified reading comprehension tests. Python.
De-Anonymizing the Bitcoin Blockchain
Unsupervised machine learning algorithms to cluster bitcoin addresses belonging to the same user. Python.
Trending Tweet Detection
Algorithms and a prediction engine for detecting soon-to-be trending tweets. Python and Tweepy.
Projects
Loupe
A native macOS photo organizer for libraries of 100,000+ RAW photos: search, rate, and tag them, then hand your picks to Lightroom. Swift, SwiftUI, and Core ML.
HabitCal
A habit and todo tracker with calendar visualization, daily, weekly, and monthly goals, and natural-language quick-add. Next.js, Express, and a native SwiftUI iOS app.
Flyover
Stitches off-the-shelf drone footage into composite maps. Python and Tensorflow.
TexIt
Webapp to convert images of handwritten assignments into LaTeX. Python, Flask, and SQL.
Amazon Kindle Product Search
End-to-end product discovery and recommendation prototype for the Kindle reader, based on the Amazon Xray NLP backbone. Java.