Production AI
Shopping AI/ML
Operate the system around the model.
I lead Shopping relevance integration and production AI/ML operations at PayPal, including model rollouts, monitoring, reliability, incident response and on-call ownership.
Krishna Pinnamaneni · AI engineering leader and author
I build AI systems
people can trust.
At PayPal, I lead engineering managers and operate production AI. Through LaunchDesk, I build agent workflows, evaluation tools and privacy-focused products. My books make that engineering discipline accessible to other builders.
- 50
- people led through
three engineering managers - 2,000+
- engineers enabled
across 100+ teams - 100+
- independent apps
across platforms - 20+
- AI books and
engineering field guides
The thread through my work
Powerful technology needs
engineering judgment.
Over 16+ years at Verizon, ADP and PayPal, I have built products, created shared platforms and developed engineering leaders. That experience shapes how I approach AI: connect the model to a real workflow, make failures visible and keep decisions accountable.
I stay close to implementation. My independent work explores the details that determine whether AI is useful: evidence behind an answer, permissions around a tool call, private data at a model boundary and evaluations before a release.
Leadership impact at PayPal
Production responsibility.
Measurable outcomes.
My current organization spans Home, Shopping, Loyalty, Profile Hub and Growth. I lead 50 people through three engineering managers; I have managed managers since 2022.
Platform adoption
2,000+ engineers
Build capabilities that teams adopt.
I created PayPal Design System / PayPal UI and scaled adoption to 100+ teams across Web, iOS and Android. Delivery KPIs showed 90% productivity improvement and 95% faster time-to-market.
Engineering quality
99.99%
Turn around delivery and quality.
I rebuilt Growth and Shopping around clear ownership, architecture reviews and automated quality controls, reducing defects by 99.99% and raising test coverage from 0% to approximately 90%.
Consumer impact
3M+ enrollments
Bring ambitious products to market.
I led engineering delivery for the PayPal+ Germany launch, reaching 3M+ enrollments in its first 30 days. The global program subsequently reached 40M enrollments.
Architecture at scale
50M+ daily requests
Create foundations that last.
I architected Subscriber Index, a GraphQL relationship platform spanning 14 APIs and five events, and created PayPal’s first React Native framework and dynamic-theming architecture.
Executive initiative
3 months
Align an organization around delivery.
As Single-Threaded Owner, I aligned 100+ teams to deliver a complete consumer-app refresh in three months. I also hired 14 engineers in under 12 months.
Selected outcomes from my PayPal work across AI operations, shared platforms, product delivery and engineering quality.
Independent AI engineering
The implementation is part of the story.
These projects are personal work. They show how I approach grounded extraction, model access, evaluation and human control.
01
Evidence and evaluation
KaiInbox
An extracted answer needs evidence.
Financial-email workflows turn receipts, bills, renewals and refunds into actions. I built source-evidence validation and a 61-case deterministic regression suite. An optional structured LLM extraction hook requires claims to match source-text anchors.
Explore KaiInbox
02
AI privacy and operational controls
Sanctuary
Protect sensitive information before a model sees it.
I implemented PII detection and policy checks before model-provider calls, tenant-scoped routing, circuit-breaker fallback and audit records, with an offline labeled corpus for detector validation. Sanctuary Browser Guard previews, redacts and blocks sensitive prompts before they leave the browser.
Explore the app listing20+ AI books within a catalog of 25+ books
Writing the engineering
playbook for AI.
I write practical field guides for engineers and technical leaders building with AI. The focus is architecture, context, evaluation and governance that can be applied to everyday engineering decisions.
A little more context
The work behind the perspective.
What connects your AI and leadership work?
At PayPal, I lead engineering managers and own Shopping AI/ML integration and production operations. Independently, I build agent workflows, evaluations and privacy tools. The same questions run through both: what should the system do, how do we know it works, and who is accountable when it fails?
Where can I explore the work?
My portfolio brings together AI case studies, 100+ apps across platforms, technical books and browser tools. It shows the engineering choices behind the projects and links to the published products and books.
Explore the full portfolio ↗Start a conversation
Let’s build dependable AI.
I welcome conversations about engineering leadership, AI product strategy, agent systems and developer platforms.