Open to Group PM · Director of Product roles
Executive Director · Agentic AI Pioneer · 0→1 Builder · Enterprise at Scale

The hardest part of AI isn't building it. It's making humans trust it enough to use it.

I'm Kaushik Patel, PhD, EMBA. I've spent the past seven years building agentic conversational AI systems for enterprise users at scale, starting in 2018 before the industry had settled on a name for it. The problems I've been working on, trust, adoption, regulated environments, are the same ones the field is focused on now.

Connect with Me See the Journey ↓
0M Platform value shipped
0M Customers reached
0+ Years in product
0 LLM capabilities shipped in 2024
Kaushik Patel, Group PM / Director of Product
Scroll
Core Competencies

What I Build Best

Agentic Conversational AI

Building agentic AI systems for enterprise users since 2018, before the industry had the terminology. Multi-turn dialogue, autonomous routing, live data retrieval, multi-step workflows. Deployed at the scale and in the regulated environments where it's hardest to earn trust.

Agentic SystemsLLMsRAGMCPConversational AI

0→1 Product Development

Founding PM for Amazon's AUSTIN global EHS platform serving 1M+ workers. Invented the Works With Alexa certification program from a data anomaly. $30 prototype → $120M platform in 13 months.

Zero-to-OneFounding PMIoTPlatform

Experimentation & A/B Testing

Data-driven product decisions are the baseline. I find the signal in noisy data, design experiments that prove causation, and ship changes before the political window closes.

A/B TestingStatistical SignificanceCausal InferenceRapid Iteration

GTM Systems & Enterprise Sales

Translated Amazon's customer obsession into JPMorgan's compliance culture. Identified the right buyer persona, built the GTM motion around operations leaders rather than IT, and made compliance an accelerant rather than a blocker.

Enterprise GTMJPMCRegulatoryBuyer Personas

PhD Technical Depth

Can speak fluently with ML engineers, data scientists, and researchers. Don't just translate, participate. This is how I earned trust with scientists, got to the real problem, and positioned fitting solutions.

PhDEMBAResearch MethodsTechnical PM

Cross-Functional Influence

Won fights against sales and R&D opposition to launch programs that mattered. Convinced VP leadership on build-vs-buy. Led teams without authority across Amazon, JPMorgan, and enterprise clients.

Executive InfluenceStakeholder MgmtChange Management
How I Lead

Leadership Philosophy

01

Prove before you ask.

Build the prototype. Run the pilot. Have the data before the meeting. Nobody can say no to results that already exist. That's how a $30 Arduino became a $120M platform in 13 months. The network expansion was approved before the political window could close.

02

Trust is the product.

Every system I've built, from AI platforms to IoT sensors to LLM capabilities, only delivers value when humans trust it enough to act on it. Technical excellence is table stakes. Earning trust from workers, sales reps, regulators, and executives is the actual job.

03

Find the signal others miss.

The Works With Alexa certification program came from a pattern buried in support tickets that nobody had connected. The best product opportunities don't announce themselves, they hide in data anomalies, customer complaints, and the gaps between what teams measure and what actually matters.

The Through-Line

Before anyone called it agentic.

The term is new. The work isn't.

In 2018, I built a conversational system for 350,000 Amazon warehouse associates to report safety issues in natural language, hands-free and mid-shift, across dozens of languages. It asked follow-up questions, routed the issue, and took action. Nobody called it agentic then. We called it a safety reporting tool.

At JPMorgan Chase, I built a production system with a conversational interface, multi-step autonomous workflows, and live data retrieval, inside one of the most regulated environments in financial services. The specifics are confidential. The pattern is the same.

The problems that make agentic AI hard in practice, earning trust from skeptical users, operating inside real constraints, measuring adoption not just deployment, are ones I've been working through for years.

2018

Conversational Safety Reporting

Agentic NLP system for 350K+ Amazon warehouse associates. Multi-turn dialogue, autonomous routing, real-time action. Sustained increase in reporting volume across the network.

AgenticConversational AI350K usersEnterprise
2022

Agentic System in Production · JPMC

Conversational interface, multi-step autonomous workflows, and live data retrieval, built inside one of the strictest regulatory environments in financial services. Specifics are confidential.

AgenticRAGLLMRegulated AI
Now

What the Industry Is Focused On

Agentic AI is where most teams are investing. The hard parts, production trust, enterprise scale, regulated environments, are the same problems I've been working on.

MCPRAGAgentic SystemsRegulated AI
Career Timeline

The Journey

Each role built toward one thesis: trust is the product. Everything else is the mechanism.

Oct 2022 – Present JPMorgan Chase Executive Director

Lead AI/ML Product · Office of the CTO

Built and shipped a production agentic system at JPMC: conversational interface, multi-step autonomous workflows, and live data retrieval, inside one of the strictest regulatory environments in financial services. Then shipped 6 LLM capabilities to 85M customers. Compliance wasn't the enemy. It was the design constraint that made the products worth trusting.

Agentic
85M
6
Jun 2020 – Oct 2022 Amazon · Fulfillment Sr Manager, Product Lead

Founding PM · Agentic Safety Reporting System

Built a conversational agentic system that let 350,000+ warehouse associates report safety issues in natural language, hands-free, mid-shift, multilingual. The system asked follow-up questions, routed autonomously, and took action without human in the loop. Reporting volume increased significantly and held there. Frontline workers trusted the AI enough to use it. The $30 Arduino prototype that proved the concept became a $120M platform 13 months later.

350K+
$30
$120M
Jul 2018 – Jun 2020 Amazon · AUSTIN Founding PM

Founding PM · AUSTIN Global EHS Platform

Amazon had 1M+ workers and a patchwork of safety systems. Ran build-vs-buy analysis, convinced VP leadership, and became founding PM of AUSTIN, Amazon's global EHS platform. Built an IoT sensor portfolio. A preventable injury happened before the platform was ready. That shaped everything.

1M+
Global
IoT
Oct 2017 – Jul 2018 Amazon · Alexa Senior PM

Alexa Smart Home · Certification & Trust

Customer trust was eroding. Pulled support ticket data, diagnosed the root cause as third-party device quality, not Alexa itself. Invented the Works With Alexa certification program from a data anomaly. Find it in the data, prove the fix, build something that earns its place.

WWA
Data
0→1
May 2017 – Oct 2017 AWS Senior PM, ML

ML Platform · Solution-Selling at Scale

AWS sales teams were product-pitching, not solution-selling. Built an ML platform using collaborative filtering to change that behavior. The real challenge wasn't the model, it was getting sales reps to trust AI recommendations enough to act on them.

ML
ColFilt
GTM
Jul 2014 – Apr 2017 Enablon PM → Client Success

Enterprise EHS · Fortune 500 Deployments

Deployed enterprise EHS platform to Fortune 500 organizations. Made complex safety tech accessible to frontline workers across language barriers, loud environments, cultural differences. Moved to Customer Success, co-designed with workers afraid of surveillance, and shifted the go-to-market from IT buyers to operations and safety leaders.

F500
GTM
Multi
By the Numbers

Measured Impact

Data is how I prove the fix before asking permission.

Platform Value Growth

From $30 prototype to $120M in 13 months

0
M
Platform value created
$30 Arduino → Amazon network-wide in 13 months
0
M
Customers reached
LLM capabilities deployed at JPMorgan Chase
0
Capabilities shipped in 2024
Regulated LLM features in production at JPMC
0
M+
Workers on AUSTIN platform
Amazon global EHS platform as founding PM
0
+
Years in product
PhD · EMBA · technical & business depth
0
K+
Workers on agentic system
Conversational AI safety reporting, Amazon fulfillment network
Projects

What I'm Building Now

Live

Calm Family

AI-powered app for families navigating separation and divorce, the hardest transition families face. Full-stack product built with Next.js, Supabase, Vercel, and the Claude API.

The pattern I've built for 15 years, making complex, high-stakes systems human, applied to the problem I care most about. This is 0→1 at its most personal.

Next.js Supabase Vercel Claude API 0→1
Visit calmfamilyapp.com →
Shipped

Works With Alexa Certification

Invented a certification program for third-party Alexa Smart Home devices from a data anomaly in support tickets. Turned a trust crisis into a quality signal.

Amazon Alexa 0→1 Data-Driven
Shipped

AUSTIN · Amazon Global EHS

Founded and built Amazon's global Environmental Health & Safety platform for 1M+ workers. Won the build-vs-buy debate at VP level. Built the IoT sensor portfolio.

Founding PM IoT Platform 1M+ Workers
Shipped

JPMC LLM Capabilities Suite

Six AI/LLM product capabilities shipped in 2024 to 85M JPMorgan Chase customers. Navigated one of the strictest regulatory environments in financial services.

LLMs FinTech Regulated AI 85M Users
Live

ProdCon · JPMC PM Community

Founded and scaled ProdCon, JPMorgan Chase's internal product management community, to over 10,000 members. Built the programming, speaker series, and knowledge-sharing infrastructure that connects PMs across the firm.

Community is a product. ProdCon is the proof.

Community Building 10K+ Members JPMC PM Leadership
What Colleagues Say

Testimonials

"Kaushik is the rarest kind of PM, one who can go deep technically with engineers, then walk into an exec review and reframe the same problem for the business. He invented the Works With Alexa program from a pattern no one else had noticed in support data. That's just how he thinks."

SR
Senior Principal, Amazon Alexa Worked together 2016–2018

"What sets Kaushik apart is his instinct to prove before asking permission. He built a $30 prototype, ran a facility pilot, and had results before anyone could say no. That combination of speed and rigor is incredibly rare, and it's why a prototype became a $120M platform."

MK
Director of Engineering, Amazon Worked together 2020–2022

"Shipping regulated AI to 85M customers is not a product challenge, it's a trust challenge. Kaushik is the person you want on that problem. He understands compliance not as a blocker but as a design constraint, and he ships anyway. We got six capabilities live in 2024."

AP
Managing Director, JPMorgan Chase Worked together 2022–present

See more recommendations on

LinkedIn Profile
Let's Talk

The problems worth working on usually find each other.

I'm looking for Group PM or Director of Product roles where AI/ML platform leadership, 0→1 experience, and PhD technical depth matter. Not a pitch, just a conversation.

Seattle, WA (open to remote)
Group PM · Director of Product
AI/ML platforms · 0→1 · Regulated AI
Send Email LinkedIn